2022-03-19 03:01:56 RT @_alialkhatib: i'm just saying i'm going to be using this in talks about how AIs manufacture their own representations of the world, and… 2022-03-19 01:36:54 RT @ahandvanish: The summary: vaccination definitely seems to reduce the risk of #LongCovid, often by 40-50%. *But* solidly 9.5%-14% of b… 2022-03-18 21:11:55 Q: What have you read on similarities, differences, overlap between dataset annotation vs. content moderation? (To me, it seems like research on each has relevance to the other) 2022-03-18 04:10:42 RT @owasow: Has anyone ever built a stats course around something like “statistical controversies”? Idea would be to have students read, en… 2022-03-17 22:42:01 @sherrying @baxterkb Thank you both for checking them out! 2022-03-17 22:41:36 RT @baxterkb: Thank you @math_rachel for this great collection! #dataEthics #AI https://t.co/RuvWBC59Vt 2022-03-17 01:31:32 RT @MosaicML: Shoutout to @jeremyphoward, @math_rachel and the whole @fastdotai team. Two-way callbacks and other ideas helped a lot in des… 2022-03-17 01:30:59 Patients can have a deep understanding of medical literature & 2022-03-16 06:16:07 In 3 months, 1136 kids in Qld were hospitalized w/ COVID-19, incl 23 in ICU This article mentions hand hygiene, but not the lack of masks, worse school plan than other states, or that CHO said "kids don't get sick" & https://t.co/OKDwC87XZb 2022-03-16 00:15:39 @csdoctorsister Congrats!! 2022-03-15 22:25:36 @starkcoffee Ha! I'm glad to hear that about buses. I went to the grocery store yesterday (North Brisbane/Moreton Bay) and < 2022-03-15 05:56:25 @Littlefishb1ue @levithatcher I also worry about the seeming uptick in strokes, cardiac events, etc being reported several months after having covid. 2022-03-15 05:40:33 There's no such thing as "not a math person" More details in my talk (hosted by @manisharoses) https://t.co/tbsOVwM6af And in my post 8/ https://t.co/sgOWpDbKPg 2022-03-15 05:06:37 “My view of mathematics is that it is an absolute mess which actively pushes out the sort of people who might make it better." -- Piper Harron (whose number theory mathematics PhD thesis from Princeton was widely covered for its creativity & https://t.co/whVMegVXiA 2022-03-15 05:04:23 "Constantly reading and hearing about underperforming Black, Latino, and Indigenous students begins to embed itself into how math teachers view these students... Teachers start to expect worse performance from certain students." 6/ https://t.co/Xcu8I8mRHA 2022-03-15 05:02:07 Imagine a world where children are not allowed to sing songs or play instruments until they reach adulthood, after spending a decade or two transcribing sheet music by hand. This is how math is often taught. A mathematician's lament, by Paul Lockhart 4/ https://t.co/Nk425LEOtM 2022-03-15 05:02:06 There's no such thing as "not a math person." On the cultural factors, misconceptions, stereotypes, & https://t.co/sgOWpDtU3o 2022-03-15 04:46:42 new post: 7 Great Lightning Talks Related to Data Science Ethics https://t.co/gdRPv1cKNU @AustralianData 2022-03-15 03:49:54 @levithatcher I'm more worried about long covid and risk of long-term organ damage. More in this thread: https://t.co/PRykhxIT0p 2022-03-15 03:46:48 @ClimateActionN8 I'm so sorry for your loss, and for the denialism that makes it even harder to process. 2022-03-15 01:11:51 It is so so alienating that govt, media & 2022-03-12 08:11:00 CAFIAC FIX 2022-01-23 06:26:33 Above quotes are all from "Re-imagining Algorithmic Fairness in India and Beyond" by Nithya Sambasivan, Erin Arnesen, @benhutchinson @tulseedoshi @vinodkpg. I recommend reading in full: https://t.co/rQ5DgLWP3Z 6/ https://t.co/Mul6hQYA8G 2022-01-23 06:26:31 significant scope creep in high-stakes AI: the tender for a facial recognition system was initially motivated by detecting criminals, later missing children, and then to arresting protestors 5/ https://t.co/vmVE2aq6X0 2022-01-23 06:26:30 - popular lending apps equate ‘good credit’ with whether user called their parents daily, had stored over 58 contacts, played car-racing games... - women overwhelmingly availed & 2022-01-23 06:26:28 Barriers to responsibility in computer systems (Nissenbaum): - the problem of many hands - bugs - blaming the computer - ownership without liability "These barriers become more complicated when technology spans cultures: more, and more remote, hands are involved..." 2/ https://t.co/IH95eeY96P 2022-01-23 06:26:26 On the importance of local context in building data models: list of data proxies in India, which can differ from data proxies in USA from "Re-imagining Algorithmic Fairness in India and Beyond": https://t.co/rQ5DgLWP3Z by Sambasivan et al 1/ https://t.co/Zl1b6R1kax 2022-01-23 05:49:31 @hspter Thank you Hilary! 2022-01-23 01:46:50 There are media & 2022-01-23 01:46:49 Thank you to everyone who continues to take covid seriously 2022-01-22 05:45:51 RT @HabenGirma: Professor @EmilyMBender tells Google & 2022-01-22 05:41:12 RT @HabenGirma: A Google VP compared AI to Deafblind people, misinterpreting #HelenKeller’s words to paint her similarities to #AI. This hu… 2022-01-22 05:34:44 RT @lisa_iannattone: @DGBassani We’ve reached the “manufacture consent for mass infection by convincing the reasonable majority that they’r… 2022-01-22 02:13:28 @the_jason_lee Thanks Jason! 2022-01-22 02:12:38 "We are constantly breathing in one another’s lung secretions." I wish more people would see poorly ventilated, unfiltered indoor air this way, so we could address it. 2022-01-22 02:08:09 I think about this article a lot. Improving indoor air would alleviate burden, disability, & https://t.co/mt9Ex546mX 2022-01-20 20:37:55 RT @ubiquity75: The point is not whether or not the chatbots have feelings. It’s the codification and normalization of abuse of women throu… 2022-01-20 19:59:03 RT @EmilBergholtz: Main pandemic lesson: More disease does not prevent future disease. A large wave does not prevent future ones -- it's t… 2022-01-20 01:45:28 Reminder: N95s can be reworn for a long time. (I'm seeing misconception that the N95s being sent out by USA govt are only enough to last for a day) https://t.co/Nvs7PyndCd 2022-01-19 23:21:52 I support waiving vaccine patents and believe everyone globally deserves access to vaccines, which reduce risk of death & At same time, I expect new variants can arise in wealthy, western countries that try to act like transmission/cases don't matter & 2022-01-19 23:21:51 In Australia, > Can people stop saying vaccines alone are enough? We need vax + ventilation + N95s + tests https://t.co/Psu36KFSSX 2022-01-19 21:43:09 Oct '21: Delta is in retreat Jan '22: Omicron is in retreat May '22: Pi is in retreat Nov '22: Rho is in retreat April '23: Sigma is in retreat Nov '23: Tau is in retreat No need to invest in ventilation upgrades, N95 distribution, & 2022-01-19 01:24:18 @mmitchell_ai I'm sorry you went through this and thank you for speaking out about it 2022-01-19 01:21:04 RT @mmitchell_ai: Companies engage in retaliation, and companies engage in smearing. Companies break labor law. None of the law really prot… 2022-01-19 01:00:31 "... owing to [AI's] general-purpose nature and ambition, it is applicable in a wide array of industries and domains, which amplifies its ability to deepen existing fault lines." 2022-01-19 01:00:30 "With or without AI, our society has deep problems related to the power of corporations, automation & 2022-01-19 00:45:21 RT @M_B_Petersen: Myth 3: Social media are echo chambers. No, research shows that, for most, social media breaks the bubble. We are more c… 2022-01-19 00:45:17 RT @M_B_Petersen: Myth 2: Social media makes people hateful. No, research suggests that online hate reflects offline frustrations that mak… 2022-01-18 22:00:36 RT @NeuralBricolage: Today is a good day as any to remember how many of us were jumpstarted by https://t.co/No0I8fQOMi - and to celebrate k… 2022-01-18 22:00:11 @NeuralBricolage @jeremyphoward Thank you Helena!!! 2022-01-18 07:12:23 And yet so many politicians (& 2022-01-18 03:34:39 @AnnastaciaMP Please update your graphic and your masks to N95/respirator. Surgical masks are not enough. Great article with more details: https://t.co/jtaNaMpzAN 2022-01-18 03:03:56 RT @UpFromTheCracks: In memory of Dr. King, take the time to read @safiyanoble’s groundbreaking essay tracing contemporary logics of digita… 2022-01-18 02:04:32 @qldhealth Please provide all Qlders (including kids) with KF95s/N95s & Qld health messaging & https://t.co/GfBU1oOYe1 2022-01-18 01:52:17 Kids vaccine centers in Qld don't require kids to wear masks The fact that parents who choose to wait in line for hours to get their kids vaxxed 1st week they are eligible, are unaware about importance of masks is a failure of public health messaging. 8/ https://t.co/lMFqp1fm6t 2022-01-17 23:30:18 - Australian Border Forces (ABF) can force people to hand over their passcodes to allow a phone search - text messages were sent & - ABF doesn't release any data on frequency https://t.co/LEbTgcYBbs 2022-01-17 09:21:09 RT @emilytroutman: Whether schools close or not isn't really my point here... absence of usable data is a form of censorship. With COVID,… 2022-01-17 08:11:00 CAFIAC FIX 2022-01-13 07:04:34 @JNilsson_lab @PhilippBayer @markagray The studies on long covid amongst the vaccinated worry me. Vaccines alone aren't enough and we should seek ventilation updates, air filters, and N95s when in crowded indoor areas https://t.co/w6AGYrXYOT 2022-01-13 04:13:56 @1chrishartley @1CeliaGreen Yes, and everyone, including kids, was required to use hand sanitizer, but kids masks were optional 2022-01-13 04:11:45 @agent_lani Yikes! The nurse offered my child a glass of water, but I quickly nixed that idea 2022-01-13 04:09:30 @NicA87363148 @1CeliaGreen Sounds like the last time I went to the pharmacist . I kept going outside while I waited & https://t.co/7BjunAym3v 2022-01-13 04:03:05 @1CeliaGreen Similarly, this was our biggest risk in a long time. A few of the staff were wearing N95s, which was an improvement from when my husband & The message about need for N95s & 2022-01-13 03:13:11 I really can't get over that the recommended government response in Australia to not-yet-eligible-to-be-vaxxed children being exposed to covid at school is to remind parents about HANDWASHING. In 2022. This is such a failure. https://t.co/y1QdI4ZimI 2022-01-13 03:03:03 Half the kids at the kids vaccine clinic were not wearing masks, and we spent 2.5 hrs waiting in line. My 6 y.o. & We'd already had 2 cancelled appts, so walk-in seemed best option https://t.co/X7YscDxAsi 2022-01-12 08:26:37 RT @joelle_emerson: Please please put your littles in good masks. An amazing resource: https://t.co/icg0LSBHZm 2022-01-12 03:44:55 @alexhanna @cephaloponderer @bernardjkoch Thank you Alex! 2022-01-12 02:08:45 @unsojo @huggingface Congrats!! 2022-01-12 02:06:46 @csdoctorsister @Kalandra67 Ha! Thank you @csdoctorsister 2022-01-12 02:04:50 "While community alignment on benchmarks & 2022-01-12 02:04:48 Advantages of benchmark datasets: - barriers to participation are reduced (everyone can use) -scientific community can easily assess progress - less reliance on slower institutions for evaluating progress 4/ https://t.co/i38SegwKEF 2022-01-12 02:04:47 Dataset usage patterns across ML: - increasing concentration on fewer & - majority of papers within most tasks use datasets originally created for other tasks, - dominant datasets have been introduced by researchers at just a handful of elite institutions 3/ https://t.co/F34tOnxP9L 2022-01-12 02:04:45 "There is increasing inequality in dataset usage globally, and > https://t.co/SqwB6BKnqs by Bernard Koch @alexhanna @cephaloponderer J Foster 2022-01-12 00:10:42 Handing out free N95s at every store, office, hospital, & Mass biometric data collection in name of "covid safety": logical, high-tech, let's do it! < 2022-01-11 23:01:12 @alicetiara @kingjen @zephoria Thank you! 2022-01-11 22:00:10 @MattAlhonte @mmitchell_ai Thanks! 2022-01-11 21:59:59 @robinberjon Thank you! 2022-01-11 08:11:00 CAFIAC FIX 2022-01-06 03:33:16 RT @julieletchner: The bad apples are not the problem. The problem is how they flourish in your department. It's you, the complicit 85%, wh… 2022-01-06 02:48:42 Shortcomings of majority vote approach (for annotations): - limits representation of minority perspectives - labels may not be internally consistent - falsely assumes that annotator judgements of diff instances are independent 3/ 2022-01-06 02:48:41 "In many subjective tasks, there often exists no such single 'right' answer and enforcing a single ground truth sacrifices the valuable nuances embedded in annotator's assessments of the stimuli and their disagreements" 2/ 2022-01-06 02:42:33 - Annotator disagreements may capture important nuances ignored by a single ground truth - A multi-task based approach yields same or better performance than aggregating labels prior to training Interesting paper: https://t.co/lQoFjUtmbS @aidaa @blahtino @vinodkpg h/t @rajiinio https://t.co/UHT7Kx73ib 2022-01-05 22:11:38 RT @NataliaHodgins: If you can’t get a PCR test and you end up w/ long-Covid, what are the implications? Well, in the US you cannot access… 2022-01-05 21:58:42 RT @MarisaKabas: people saying that you need to accept that everyone's going to get covid at some point have never dealt with chronic illne… 2022-01-05 21:15:48 @dubikan @megyoung0 @MikeKatell While I agree we should be clear about what is ML vs. not, there are properties automated systems have in common, which are distinct from non-automated systems. This is about ML, but most points apply to automated systems too: https://t.co/fAhRTlS37N 2022-01-05 08:00:15 @dubikan @megyoung0 @MikeKatell The paper highlights that most ML practitioners consider OCR to be ML (I agree-- neural nets have long achieved best performance on OCR, so most used in practice). Key point of the paper is about the *difference* in views between policymakers vs. ML practitioners though 1/ 2022-01-04 23:46:09 "Defining AI in Policy vs Practice" by PM Krafft, @megyoung0 @MikeKatell K Huang, G Bugingo The gap in AI definitions may result in regulatory efforts overemphasizing concern about future tech at the expense of pressing issues with existing deployed tech https://t.co/Zyh7SDw6m9 https://t.co/6iFbzvpxDV 2022-01-04 23:46:08 Analysis of 40 policy documents found a much greater focus on human thinking/humanoid definitions of AI. Focusing on hypothetical systems not-yet-created raises risk of ignoring autonomous systems already in place. 2/ https://t.co/TMeAUhzpq7 2022-01-04 23:46:06 Policymakers & https://t.co/Zyh7SDw6m9 PM Krafft @megyoung0 @MikeKatell et al 1/ https://t.co/1YUSWIhhBt 2022-01-04 23:10:42 @NinaScience This is good to see! 2022-01-04 23:10:11 RT @NinaScience: @math_rachel Safe Work Australia is trying to help people understand the role of indoor air (Australian government statut… 2022-01-04 22:48:52 RT @QUTDataScience: Congratulations to our Centre's Data Scientist in Residence, Prof Rachel Thomas! 2022-01-04 22:38:53 RT @timnitGebru: I personally would rather take misogynistic & 2022-01-04 21:58:51 @ruthstarkman @BostonReview Oh wow, thank you! I'm honored . Happy 2022 to you too! 2022-01-04 06:01:59 To clarify, the virus will keep mutating and people aren't forming lasting immunity https://t.co/9Mo1na1yqq 2022-01-04 03:34:36 @auxiarch And the shortest of short-term thinking (unable to consider outcomes even 3-6 months from now) 2022-01-04 02:30:42 I used to think that countries (leaders) were unwilling to learn from each other's mistakes. Now, I realize that they aren't even able to learn from their own mistakes. https://t.co/htgbSK5sLX 2022-01-04 02:21:52 @wjoelburton @WomeninAIEthics Thanks Joel 2022-01-04 00:37:05 Independent group of scientists & https://t.co/myeyocBqYD https://t.co/7NCwBPmuer 2022-01-04 00:37:03 Australia largely went all-in on vaccines alone: - 91% of age 16+ have 2 doses (impressive!), but only ~12% have 3 doses (what you need) - most older people received AstraZeneca - won’t START vaxxing kids 5-11 until next week (w/ 8 week dose spacing) 2/ 2022-01-04 00:37:02 For those outside Australia , some covid context here: - Major strategy pivot towards “herd immunity” - ⅓ of cases in last 2 years occurred in last 4 days - this is an underestimate, since massive testing shortage (25% test positivity rate) 1/ https://t.co/vi9w01TTmV 2022-01-04 00:12:18 Thank you @WomeninAIEthics ! I am in great company. https://t.co/mjsFLfHZfH 2022-01-03 22:46:35 @SoulSoldSeprtly They were until recently, but now much of Australia has pivoted to "herd immunity". I spent the 1st year of pandemic in USA, and am horrified to see Australia now make such a 180 in their covid approach https://t.co/MilfLdF8Pz 2022-01-03 22:36:55 My state is home to world-class expert Prof Lidia Morawaska (one of Time's 100 most influential people for her work on #COVIDisAirborne), yet she isn't even being consulted on how can we can make indoor spaces safer here. 7/ https://t.co/3yt0esHczv https://t.co/TfMkDGl9Tp 2022-01-03 22:35:14 We need the public to understand #COVIDisAirborne, but how can they, when this is the messaging they receive? 6/ 2022-01-03 22:34:51 At my local pharmacy, one of the pharmacists removes her mask whenever she is > 2022-01-03 22:34:15 Ideas about 1.5m (or 6 ft) being safe were largely based on droplet transmission, but it has long been clear that covid is airborne. Yes, distance helps, but ventilation, air filters, quality masks, indoors vs. outdoors are all crucial too. See 4/ https://t.co/Yyj2ikIYKy 2022-01-03 22:33:01 @WHO @ausgov @QldGov The Chief Health Officer for my state told people on Jan 1, 2022 that it’s acceptable to take your mask off indoors, in a group work meeting, if you are 1.5 meters away from others. 3/ https://t.co/Q5mDdpmlpr 2022-01-03 22:32:28 I live in a major city (pop: 2.5 million) and these are signs I saw RECENTLY (Nov-Dec 2021) about covid, including at my local hospital & NONE mentioned masks, ventilation, or that #COVIDisAirborne. 2/ https://t.co/qZnXizP6Tc 2022-01-03 22:28:06 Understanding #COVIDisAirborne is crucial, but I'm alarmed that Australian politicians (and the medical doctors/pundits advising them) are not aware of this. 1/ https://t.co/Hfb7suVUb3 2022-01-03 19:49:18 @radekosmulski @fastdotai Congrats Radek! 2022-01-03 10:29:28 RT @adamhamdy: 9. The history of public health is littered with examples of vested interests resisting calls for better sanitation, better… 2022-01-03 10:28:41 RT @adamhamdy: 5. This is a tired tactic of vested interests throughout history. Smoking? They're destroying your freedom of choice. Occupa… 2022-01-03 07:31:05 @ceruleanot @bennessb I found this article helpful: https://t.co/ray74fCLFF 2022-01-02 23:33:45 RT @yaneerbaryam: Addition: the genetic combinations of Omicron and Delta that form when someone is infected by both can be expected to res… 2022-01-02 05:06:22 RT @pookleblinky: By broadcasting mutually contradictory, needlessly convoluted, nonsensical policies, an institution can ensure that for e… 2022-01-02 01:16:32 @CrabbBrendan Thank you! 2022-01-02 01:16:24 RT @CrabbBrendan: Fabulous thread, so much sound information. And the actions below are spot on. #COVID19Aus. Thanks @math_rachel 2022-01-02 00:14:49 More on the importance of cleaning up indoor air (& https://t.co/uQkkor2Ztd 2022-01-02 00:00:45 @BernieDogs4 @jfeldman_epi Agreed. Also, many healthy people fundamentally do not understand the nature of chronic illness at all. (from march 2020) https://t.co/d7RmR2IjqM 2022-01-01 22:36:49 @ahandvanish @jeremyphoward Thank you Hannah, I admire your advocacy work & 2022-01-01 21:25:31 @ecq023 @jljcolorado In particular: https://t.co/tnkgUf2ONd 2022-01-01 21:24:35 @ecq023 The 6 ft rule is outdated. Distance helps, but ventilation, size of space, & https://t.co/uOao6ZyWlI 2022-01-01 18:45:36 Related: & https://t.co/L3Mzx9Ntue 2022-01-01 18:40:15 Related: on why vaccines alone are not enough & https://t.co/sVGVKQwiKn 2022-01-01 08:52:16 @EconBerger @hspter @mattmoehr @can "strictness" doesn't take into account the efficacy of the measures 2022-01-01 08:46:58 Here are some ideas about additional things we could be doing to slow the spread of omicron 12/ https://t.co/5OOW7eUeCj 2022-01-01 08:41:44 @EconBerger @hspter @mattmoehr @can Spring 2020: we were bleaching groceries & Now: much more widely known the importance of N95s/KF94s/etc (& 2022-01-01 08:29:02 @bazscott @Michael_b_evans 2022-01-01 08:27:14 Dr. John Gerrard, described as "one of Australia's leading infectious disease specialists", said at a press conference TODAY (Jan 2022) that it's fine to take your mask off indoors in a group meeting if you are spaced 1.5m apart, basically denying #COVIDisAirborne. Relevant : https://t.co/11a2VpAG13 2022-01-01 08:04:25 @infinite8horizo @ItsBouquet I can't understand it, at all. I previously felt positively about Paluszczuk, but how can she accept someone so dangerous & 2022-01-01 08:02:20 @Michael_b_evans @bazscott I'm not even going to click 2022-01-01 05:59:48 It is ableist to ignore the many people (including some children) who are high health risk and more likely to die or experience extreme illness if they get covid. Again, any reduction in transmission can save lives. 11/ https://t.co/wlk8x8Vn8y 2022-01-01 05:48:11 Due to bad messaging, many people misunderstand: - Severity & - Airborne spread Many may voluntarily choose to change their behavior if they better understood (particularly if given options: free N95s, work-from-home when possible, etc) 9/ 2022-01-01 05:48:10 Even if EVERYONE does get covid, delaying WHEN we do so has benefits. The longer we can wait, the greater the chance of new treatments being developed, more effective vaccines, better anti-virals, scaling production of paxlovid, etc. 7/ 2022-01-01 05:48:09 Many seem to be assuming that covid provides lasting immunity, even though we already know it does not. Each covid infection raises the risk of creating a new "pre-existing condition" that will make subsequent infections more dangerous 5/ https://t.co/aIeQyDkEy3 2022-01-01 05:48:08 Not everyone is going to get omicron. And the difference between 80% vs 60% (both still very big) of a population catching it would be a huge difference in absolute numbers of people that wind up permanently disabled. 3/ 2022-01-01 05:48:07 “Covid zero is impossible with omicron, so what’s the point in even trying to reduce/slow transmission?” My answer: 1/ 2022-01-01 02:46:48 @RoboticsAINews It's depressing & 2022-01-01 02:39:55 @DocSlinky Me either. Has anyone started a #GerrardMustGo campaign or similar? 2022-01-01 02:21:13 Also, we haven't even started vaccinating kids under 12 yet here. https://t.co/fR9K4nFn1B 2022-01-01 02:19:41 This is a MEDICAL DOCTOR and GOVT OFFICIAL (chief health officer) giving deadly misinformation at a press conference. #CovidIsAirborne https://t.co/2vmTMdOwVI 2022-01-01 01:55:59 @bazscott That is good to hear, thanks! 2022-01-01 01:51:51 Have any of the Australian Premiers or Chief Health Officers even publicly mentioned LongCovid yet? https://t.co/soInMZU8XP 2022-01-01 01:20:09 @qldhealth It would be great to see @QldGov provide all Queenslanders with free N95s. Also, they can be reworn many times (which reduces cost) 2/ https://t.co/CXs4EfNWfL 2022-01-01 01:18:07 @qldhealth It's important to let people know the importance of well-fitting, high-quality masks (N95s/KF94s) instead of poorly fitting surgical or cloth ones 1/ https://t.co/alMejmtLdM 2021-12-31 23:15:36 Covid wild-type: original strain, earliest pandemic waves Alpha variant: first detected in UK in Nov 2020 It's a pedantic point, but I see many people confusing wild-type & 2021-12-31 21:47:05 @juliaferraioli Oh no, I'm so sorry to hear this and angry about the factors that led to it 2021-12-31 20:57:49 @migueldeicaza @mmitchell_ai Thanks Miguel, happy new year!! 2021-12-31 20:56:19 RT @GaelVaroquaux: . @math_rachel's excellent article on pitfalls of machine learning for medicine: - Data capture systematic errors/biases… 2021-12-31 04:46:42 Check out the replies below for several good articles & 2021-12-31 02:43:33 Things can go disastrously wrong in data science projects wrong when we: - undervalue data work - ignore the context of data - neglect the role of humans - don't listen to the people most impacted 8/ https://t.co/NwY9taqzoi 2021-12-31 02:43:32 You CAN and SHOULD get involved with AI, keynote at Brisbane @women_in_ai launch @QLD_AI_Hub 7/ https://t.co/dmOD4JU2wx https://t.co/HXzNWhBsrx 2021-12-31 02:43:31 "Deep Learning is Overhyped" is Overhyped, together with @jeremyphoward 6/ @QUTCDE https://t.co/vnUcKpwp2M https://t.co/doac6DPYWL 2021-12-31 02:43:29 Fairness, Explainability, and Diverse Datasets Are Not Enough My talk at @QUTDataScience Responsible Data Science Symposium 5/ https://t.co/zBeg9GyGeb https://t.co/msZj4yYa5n 2021-12-31 02:43:27 11 Short Videos About AI Ethics 4/ https://t.co/U8gSlUxdeG https://t.co/r1tebaVSHm 2021-12-31 02:43:25 Machine Learning + Medicine: - Data can be biased, missing, incorrect - Expertise of patients overlooked - Med system is already disempowering & - Machine learning systems can centralize power Jan 2021 @BostonReview 2/ https://t.co/iIZDT1nOlh 2021-12-31 02:43:24 I want to share some essays I wrote & 2021-12-30 22:20:39 @manisharoses @jeremyphoward 2021-12-30 20:59:42 @Hinnerk_hero It's not observable through a biomarker/blood test (yet), but it can be observed through patients' reporting of symptoms. Consider how adult MDs w/ long covid are being dismissed/disbelieved, children are likely facing far more dismissal: https://t.co/pAClcdPTRV 2021-12-30 20:46:42 @deingaraus @hipsterelectron I know, it's infuriating. 2021-12-30 20:45:33 @Hinnerk_hero I think considering the absolute numbers of kids impacted in UK & 2021-12-30 20:35:33 @Hinnerk_hero Unfortunately the meta-reviews of long covid in kids that I've seen had serious methodological issues (failed to take into account high false negative of PCRs in kids, seroreversion, long covid correlated w/ failure to seoconvert, etc) More here: https://t.co/bg5d7STs23 2021-12-30 15:22:59 @againstempiric1 Slowing down/delaying infections gives us more time for new treatments or vaccines to be developed. Also, each subsequent covid re-infection carries additional risk (some people are fine 1st time but not 2nd): https://t.co/vILMWkReOw 2021-12-30 15:18:24 @againstempiric1 There is still huge value in reducing transmission & https://t.co/exgtZPitYN 2021-12-30 09:49:39 @kent_boogaart No problem, thanks! 2021-12-30 09:37:17 @danielmurfet Thank you, I'm glad they were helpful! 2021-12-30 09:36:06 @kent_boogaart Thank you. I missed that! Do you have a link about the announcement? 2021-12-30 09:21:17 @MeganEDoherty @edyong209 Thank you for writing it! 2021-12-30 09:20:47 @r_prior @edyong209 Thanks, I had missed this! 2021-12-30 09:20:20 @loscharlos @edyong209 Thank you! 2021-12-30 09:20:12 @talirichards4 Thank you! 2021-12-30 08:19:37 RT @profesterman: Long Covid is the elephant in the room, but it seems invisible to our politicians | Adrian Esterman https://t.co/EbIUdDWe… 2021-12-30 07:58:43 It was magical to get 8 months of safety in Qld, and now it is horrifying to watch our leaders throw it away w/ false fatalism. Leaders are ignoring existence of children, high-risk chronically ill & https://t.co/MqwaNoyM9Q 2021-12-30 07:58:42 Previously, my partner and I went for 13 months of pandemic in the USA with no in-person school and no in-person childcare. (A sincere thank you to @manisharoses, who changed how we think about education!) 6/ https://t.co/f561WSYMqi 2021-12-30 07:58:41 I believe it is possible to do in-person school safely, with mask mandates, air quality standards & https://t.co/ekYuwevQCL 2021-12-30 07:58:40 Research shows covid can cause changes to the brain, blood vessels, & https://t.co/IGR0ChWq6J 2021-12-30 07:58:39 My state has pivoted to herd immunity (saying widespread covid is not just inevitable but “necessary”) BEFORE they have even STARTED vaccinating 5-11 year olds. I will be taking my child out of in-person school. #auspol 1/ 2021-12-30 04:05:56 Question: what is the best article for a general audience on LongCovid in CHILDREN that you've seen? Ideally something similar to @edyong209's long covid coverage (scientifically rich, accessible, centering patients) in Atlantic, only specifically about kids. #LongCovidKids 2021-12-29 22:48:15 @Kyra_Yee I'm sorry your are going through this. I hope you are able to find some answers & 2021-12-29 10:53:26 @qldhealth Here is an article on what "mild" cases can include. #LongCovid has left many in constant pain & https://t.co/Rho3kakICr 2021-12-29 10:48:14 @qldhealth Initially "mild" covid cases can result in long-term damage to the immune, vascular, & Vaccination reduces risk, but significant numbers of vaxxed still get #LongCovid. Please recommend vax + N95s + ventilation + air filters https://t.co/BHGjuHSN7y 2021-12-29 03:37:24 @LongDesertTrain Thank you Ryan 2021-12-28 02:21:55 RT @MagisterBracey: I would like to formally invite financially comfortable white parents to stop speaking for Black parents on the issue o… 2021-12-28 00:22:41 RT @mmitchell_ai: I'm working w my fellow Washingtonians @cherthedev and @chelseyglasson, following the CA work of @erikashimizu and @Ifeom… 2021-12-27 10:30:18 RT @wsbgnl: The conversation that is needed is — what measures must be in place in every school to meet a minimum safety standard? Ventilat… 2021-12-27 08:20:00 CAFIAC FIX 2021-12-21 08:20:52 Harmful false binaries: - covid cases either die OR recover - our only options are lockdowns OR no mitigations at all - covid zero OR there's no point reducing transmission - you think vaccines are *all* we need OR you are anti-vaccine 2021-12-21 05:37:35 RT @NjbBari3: Have new autoimmune issues. New cognitive issues Have vascular problems like strokes other organ damage like kidney failu… 2021-12-21 05:37:32 RT @NjbBari3: So once a year, a certain proportion of all human beings, except those in countries with good COVID control will Die Hav… 2021-12-21 02:51:16 RT @negar_rz: @alexhanna @emilymbender Adding to that the experiences of people with disability in tech industry, @math_rachel has amazing… 2021-12-21 02:50:56 RT @jenbrea: @mattyglesias In fact, we aren’t even talking about transmission reduction anymore! Most of the focus is severe disease preven… 2021-12-21 02:50:41 Even if you believe it's "inevitable" everyone will get covid eventually, it's in our best interest to minimize # of times and to delay (re)infections (w/ hope better vaccines or treatments are developed). Reducing transmission is still a valuable goal. up & 2021-12-20 09:52:29 @negar_rz @alexhanna @emilymbender Thank you Negar 2021-12-19 09:34:07 The focus on vaccines alone is convenient for helping governments shift blame away from their own policy failures & https://t.co/KYVsBKpT9r 2021-12-19 08:12:13 @ProfEmilyOster Thread on the problems of only emphasizing vaccines: https://t.co/Q8K123sd5T 2021-12-19 07:57:39 RT @GreggJLevine: Another v.good thread on why the “pandemic of the unvaccinated” strategy/messaging is insufficient & 2021-12-19 06:00:36 RT @math_rachel: The last 20 months has been such a FUCK YOU to those of us with chronic illnesses: - openly eugenic sentiments (eg death n… 2021-12-19 01:11:27 Here are some things you could try @qldhealth @AnnastaciaMP https://t.co/l1NBxBq8fh 2021-12-19 01:11:26 Everywhere I go, I see signs urging me to wash my hands to stop the spread of covid. You can't even be bothered to update your signs @qldhealth??? https://t.co/pYThdPDs7Q 2021-12-19 01:11:25 The Chief Health Officer for my state just described pervasive spread of covid as "inevitable" and "necessary" IT IS ONLY INEVITABLE IF YOU DECIDE IT IS INEVITABLE!!! @qldhealth https://t.co/3MKX1vRxjP https://t.co/y29TUhfsr8 2021-12-19 00:50:57 RT @KashPrime: Getting Omicron is NOT inevitable. I hate that our leaders have essentially given up, before implementing proper widespread… 2021-12-19 00:45:26 @jenbrea @ProfEmilyOster @danaparish @ahandvanish @jeremyphoward I put together a thread on the problems with focusing on vaccines alone: https://t.co/Iadjvab4j3 2021-12-19 00:43:39 My mistake in tweet #2. Meant to QT this chart about Denmark: https://t.co/sWFSiN5HeX 2021-12-19 00:41:27 Vaccination is important, but insufficient. We must improve ventilation and wear high-quality, well-fitting masks indoors in public places. Policy makers need to take into account the whole picture 19/ https://t.co/MqwaNoyM9Q 2021-12-19 00:41:26 As @jeremyphoward wrote, we shouldn’t confuse all vaccine hesitancy with extreme anti-vaxxers. We should continue to reach out to the hesitant and to those facing logistical barriers 18/ https://t.co/Jc9D0Fd0vC 2021-12-19 00:41:25 To be clear, vaccines are incredibly helpful (just not sufficient on their own) and we should continue to address misunderstandings & https://t.co/fDQqTS2qYL 2021-12-19 00:41:24 Many leaders treat masks as punishment. There is no evidence that a “reward” of no masks is effective at motivating vaccination. For most of us, the reward for getting vaxxed is reducing risk to ourselves, our loved ones, & https://t.co/U63fgprUyP 2021-12-19 00:41:23 Many leaders assume they must prevent people from “panicking”. But it is a myth that people panic in disasters. In fact the opposite: research shows many behave altruistically when facing disasters & https://t.co/a5nRPSG88Z 2021-12-19 00:41:22 A lot of public health messaging has been patronizing, infantilizing, & https://t.co/v3xqOxloHo 2021-12-19 00:41:21 Also, we are facing a coordinated campaign to downplay the risks of covid, using strategies straight out of the climate denial playbook & https://t.co/OjqRVMdpco 2021-12-19 00:41:20 In the West, our covid response has been hobbled by neoliberal/libertarian ideas of individual responsibility, which are insufficient for collective problems (such as pandemics or climate change) 9/ https://t.co/MFCItzFYDJ 2021-12-19 00:41:19 Leaders set people up for anger & https://t.co/NJQlK2OYAG 2021-12-19 00:41:18 Covid & https://t.co/mt9Ex5l9oX 2021-12-19 00:41:17 “But people are fatigued and won’t tolerate any more restrictions.” Good news, here are a bunch of things governments could do that aren’t individual restrictions 4/ https://t.co/l1NBxBq8fh 2021-12-19 00:41:16 If you are 3x vaxxed, you should still care: - Breakthroughs not rare - Significant # of breakthroughs get LongCovid - You may need functioning hospital system - Many are ineligible for vax (kids) or can’t mount antibody response (eg cancer patients) 3/ https://t.co/ray74fCLFF 2021-12-19 00:41:15 Focusing on vaccines ALONE: - Does not work (many countries tried & - Hospitals can still collapse - Sig # of vaxxed get LongCOVID - Rooted in neoliberalism: individual responsibility as soln to collective problems We need: vaccines + N95s + ventilation + air filters 1/ 2021-12-18 10:21:20 @violinwitch Yikes, that is particularly bad. I've recently started taking photos of the wash-hands-to-fight-covid signs I see everywhere. This collage is just a subset: https://t.co/jUp6Y3f5eZ 2021-12-18 10:16:46 @JenniferMW @moniquedhooghe We do have 2 years of incorrect & 2021-12-18 06:46:45 @mmitchell_ai No problem! Also, it wasn't at the door, but for the super spreader holiday party in Norway (where at least 100 out of 120 got omicron), everyone was vaxxed and did rapid tests the day before. 2021-12-18 06:36:32 @mmitchell_ai Here is an account of vaccinated people that did rapid tests at door and at least 7 out of 11 got covid (with 2 more symptomatic & 2021-12-18 04:54:06 RT @math_rachel: More things govt could do that are NOT individual restrictions: Public education: KF94/N95 > CO2 mon… 2021-12-18 04:54:02 RT @math_rachel: Things govt could do that are NOT individual restrictions: Mail everyone free KF94s & Fund ventilation upgrade… 2021-12-18 03:27:30 @justthefacts85 @WHCOVIDResponse I don't think I'm up for it 2021-12-18 03:10:00 @jljcolorado @WHO My local vaccine hub (set up just for covid) has signs from @WHO on the 8 step process for proper hand-sanitizer use, but no signs about masks or airborne transmission. https://t.co/znca7kYxqE 2021-12-18 02:13:15 Reminder: Best practice contact tracing includes ensuring that people have enough food, security, and ability to care for their children & 2021-12-18 02:05:21 Even if leaders persist w/ false fatalism, at least be honest: "Everyone is going to get Omicron BECAUSE we are failing to act. Covid is airborne. You breathe unsafe air at work, your kids breathe unsafe air at school. Many of you will be permanently & 2021-12-18 02:05:20 More things govt could do that are NOT individual restrictions: Public education: KF94/N95 > CO2 monitors for all indoor spaces, results displayed clearly for all occupants to see Replace all the "fight covid w/ hand-washing signs" 3/ 2021-12-18 02:05:19 Many leaders say people are fed up w/ restrictions, so we can’t try to do anything about Omicron: - Premise questionable, as several polls show majority want mask mandates - There is a TON government could do (but isn’t) that does NOT involve individual restrictions 1/ 2021-12-17 22:47:22 CDC director says omicron cases doubling every 2 days, but provides no context & 2021-12-17 22:47:21 Once again, I can find more accurate info about omicron on twitter than by listening to CDC Director on major network television. 1/ https://t.co/WKF2MwJ6zh 2021-12-17 22:09:15 This whole thread. So much public health messaging is both misleading AND ineffective. The focus exclusively on vaccines (leaving out high-quality masks, ventilation, air filters) and on hospitalizations/deaths (leaving out #LongCOVID, which is far greater risk) is NOT working. https://t.co/0kZlu2XMBO 2021-12-17 21:35:02 Great thread from @rajiinio with 10 papers on different aspects of dataset-focused Machine Learning research. Updating my reading list now. https://t.co/wIYOG5S7D8 2021-12-17 19:59:54 RT @LongDesertTrain: 1/ Can't put it better than @micah_arsham does here. The CDC treats the public like 1st-graders. Their job is to infor… 2021-12-17 02:30:45 Perhaps my excitement was premature. Really, they should be requiring & 2021-12-17 01:35:45 @qldhealth Thank you, I am relieved to hear this! It would be great to accompany this with updated public health messaging #COVIDisAirborne https://t.co/SmpLcxDlbz 2021-12-17 01:05:36 Excellent news!! Thank you @qldhealth https://t.co/n14jXt4XOC 2021-12-16 22:52:14 Throughout pandemic, maxed out hospitals have led to: delayed cancer treatments, cancelled brain surgeries & https://t.co/Js3aZBfXg0 2021-12-16 08:12:45 RT @ahandvanish: Thread with articles on #LongCovid from 100 different countries around the world. This is a global problem. 2021-12-16 07:33:03 RT @jenbrea: I am just going to keep saying it. This is the largest mass disabling event in our history. The more I sit with that reality… 2021-12-16 05:13:53 @jeremyphoward @WHO I also like @RealOzSAGE graphic on making indoor air safer. Workers should not be forced into workplaces with unsafe air, nor should children be forced into schools with unsafe air. Let's set indoor air quality standards & 2021-12-16 05:04:09 @Michael_b_evans @jeremyphoward Yes, we got Pfizer for all 3. And I definitely agree that the cost for the average person is too high (and long-term it is in the government's interest to protect public health) 2021-12-16 05:01:10 @jeremyphoward @WHO The "5 Ways to Fight an Airborne Virus" graphic would make for great messaging in a vaccine clinic (& https://t.co/NfsN3NyHXt 2021-12-16 04:55:29 @jeremyphoward Our vaccine clinic had @WHO signs on proper hand-sanitizer use. What we need are signs about MASKS: WHICH ones: KF94/N95/FFP2 (everyone there except us had surgical or cloth masks) HOW: many folks weren't covering noses WHY: b/c #COVIDisAirborne & 2021-12-16 04:42:40 I'm glad @jeremyphoward & - Govt funding for ventilation upgrades & - FREE N95s/KF94s & - public education: #COVIDisAirborne - public education: #LongCovid is hell https://t.co/C0iFnJYi5D 2021-12-15 21:25:20 @cristi_vicas Yes, from human compatible 2021-12-15 20:58:56 @tedmasterweb @JanelleCShane Agree, Janelle is awesome! 2021-12-15 08:20:17 RT @jljcolorado: 28/ As some folks are pointing out, in more litigious societies like the US, part of the costs take the form of legal liab… 2021-12-15 08:17:15 RT @jljcolorado: 20/ More importantly, governments & 2021-12-15 06:39:41 Abstract, futuristic, or science-fiction-inspired images of AI (eg humanoid robots, glowing brains, & A project for alternatives: https://t.co/eQ31nLyZwE 2021-12-15 03:55:22 RT @wimlds: Is there anyone in our community who has #LongCovid? @patientled is looking for someone with #MachineLearning expertise an… 2021-12-15 02:35:28 The above notes & https://t.co/1EoYLBfYS7 2021-12-15 02:35:27 Google's adtech auction mechanism ultimately bids up their own share Uber uses theory of driver autonomy to push workers further away from capital Ultimate contradiction: superlative information gathering of the market ends up looking more like a control structure @jakusg 6/ 2021-12-15 02:35:26 Multi-sided platforms are not neutral intermediaries. Arguably, they are not even really markets. Instead, they are tools of world building: their designs & 2021-12-15 02:35:24 Instead of a commitment to individual autonomy, human rationality becomes a variable to be tuned as algorithmic mechanism designers search for desirable patterns of actions & 2021-12-15 02:35:22 In privately controlled & 2021-12-15 02:35:21 The point of an auction is to incentivize truthful disclosures of preference & 2021-12-15 02:35:19 Algorithmic mechanism design (eg ad tech auctions such as Google, multi-sided platforms such as Uber) contains an inherent contradiction: - theory of preserving autonomy & - reality of opaquely generating & @jakusg 1/ 2021-12-15 01:03:55 @FrankPasquale @in4dmatics @WeCompetition @sociotiose Thank you Frank! I'm a fan of your work and really appreciate your saying this. You made so many great points during the discussion & 2021-12-15 00:58:31 A single tool is not enough against omicron. We need vax AND well-fitting masks AND ventilation AND rapid tests AND air filters It is not too late to start setting policy goals around getting R < 2021-12-15 00:21:21 I really enjoyed this discussion with @FrankPasquale @in4dmatics @WeCompetition @sociotiose! It was an honor to be part of the Political Economy of Reinforcement Learning Systems workshop #NeurIPS2021 https://t.co/ED4nSBnR93 2021-12-14 09:43:53 Russell proposes a formally defined assistance game (Reinforcement Learning ++) that allows uncertainty in the reward function and where optimal solutions: defer to humans & 2021-12-14 09:43:51 "We cannot specify rewards completely and correctly. Our 3rd wish (reward function) is to undo our first 2 wishes (reward functions)." -- Stuart Russell #NeurIPS2021 PERLS workshop https://t.co/hlC6VY0HgA 2021-12-14 09:12:36 RT @clancynewyork: Another great thread from Rachel Thomas thinking through WHY it's still easier to find timely accurate info about Covid… 2021-12-14 09:11:58 RT @dlowd: I’d love to see Google, Amazon, and Facebook support safe AI via labor protections and antitrust measures. 2021-12-14 06:43:28 We can measure our effectiveness in how often we disagree with one another in our collaborative work. Disagreement gets at the goal of compromise. Collectively & 2021-12-14 06:43:27 The challenge is that not all data can be measured mathematically or captured as data points. Discretely measuring & 2021-12-14 06:43:26 Data is power. Because data has become so powerful, we must transfer the tools of data collection, aggregation, & 2021-12-14 06:43:25 How can we think about building systems not for individual end users, but for groups and communities? So much of computer science is limited by the assumption that there is an individual end-user who should be the benefactor of what we build. - @marylgray keynote #NeurIPS2021 1/ 2021-12-14 06:16:41 @PolyVinci Yes, many people responsible for these failures are still in leadership positions, still setting policy, still interviewed in the press, etc. 2021-12-14 05:29:52 Pre-pandemic, I knew that doctors often disbelieve patients and that medical system can be error-prone, hostile to patients, & essay I wrote in Jan 2019: https://t.co/71pnZRHysE 25/ https://t.co/ADhOKFY8VL 2021-12-14 04:46:05 Contrary to big tech execs’ cold-war style rhetoric about an arms race, what truly stifles innovation is the current arrangement where a few people build harmful tech & 2021-12-14 04:44:06 "When people ask what regulations need to be in place to safeguard us from the unsafe uses of AI we’ve been seeing, I always start with labor protections and antitrust measures. I can tell that some people find that answer disappointing..." @timnitGebru https://t.co/qCCk0gpcok 2021-12-14 02:02:56 RT @p2173: Perfect example of how "ethics" as a frame has been captured by the tech industry. Print ad in the NY Times Dec 13 2021 https://… 2021-12-14 01:52:12 @clancynewyork Thank you Eileen! 2021-12-14 01:49:38 @AtomicAdMan Thank you! 2021-12-14 01:29:36 @AtomicAdMan Yes, that has been an ongoing source of confusion/bad advice. (Although as the virus evolves to be more transmissible, the public does need higher quality masks than before) https://t.co/vrjJnOPWs2 2021-12-14 01:09:29 Dec 2021: Gatekeeping & https://t.co/fujzD9Ziyk 2021-12-14 01:09:28 It is disturbing to watch doctors share misinformation while advising the government, writing for mainstream media pubs, being quoted by journalists, etc. 23/ https://t.co/Sffd6igIpr 2021-12-14 01:09:27 Nov 2021: There is often more focus on whether the public trusts key institutions, than on how those institutions have eroded our trust (& https://t.co/zT87eVlyod 2021-12-14 01:09:26 Oct 2021: Pre-pandemic, I had been frustrated by how disagreeing with doctors is painted as inherently “anti-science”, given how often doctors disbelieve patients with chronic illness (like me). The pandemic heightened this. 20/ https://t.co/dDIPfu2Ayl 2021-12-14 01:09:25 Medicine is inherently political: which hypotheses get funded, whose expertise is recognized, whose suffering is counted. 19/ https://t.co/3E7Sy8hBGj 2021-12-14 01:09:24 Sept 2021: Reasons why western countries believe in their own exceptionalism and refuse to learn from other countries’ experiences, even as they keep making the same mistakes 17/ https://t.co/htgbSK5sLX 2021-12-14 01:09:23 Oct 2020: The psychology of many people not being able to recognize their own vulnerability is significant: believing that they will never be disabled or chronically ill, that they won't be impacted, etc 15/ https://t.co/plvbKiQr0p 2021-12-14 01:09:22 July 2020: Articulating how patients have crucial medical expertise, distinct from doctors. I shared this (+ some of my personal story) in my @StanfordAIMI keynote 14/ https://t.co/ehN43HL65R 2021-12-14 01:09:21 Late April 2020: thread of quotes from a great article by @edyong209 on why the pandemic was hard to understand 12/ https://t.co/bfMIXsW5rK 2021-12-14 01:09:20 Only later did I see how emphasizing hand-washing aligns with an ideology of personal responsibility as the solution to systemic problems. Improving the quality of our shared indoor air require a more collective response 11/ https://t.co/MFCItzFYDJ 2021-12-14 01:09:19 April 2020: I listed the overemphasis on randomized control trials & https://t.co/A8ICUbljB7 2021-12-14 01:09:18 Many voices that have been excluded from prominent media & https://t.co/TP2dnP5F7j 2021-12-14 01:09:17 While social media is rife with disinfo, I was able to find higher quality covid information (and sooner) on Twitter than what friends/family found in mainstream media. I still consider this to be the case. 7/ https://t.co/0Uh0AKYAQt 2021-12-14 01:09:16 And these failures are ongoing. of of my various attempts over the past 22 months to make sense of these failures. 5/ 2021-12-14 01:09:15 Our institutions failed: saying masks don’t work 2021-12-14 01:09:14 I went from giving keynote talks warning about disinformation (2019) to helping @jeremyphoward make multiple YouTube videos contradicting the CDC & 2021-12-14 01:09:13 22 months in, and I still can find more accurate info about Covid on Twitter than if I had to instead rely on articles in NYTimes by Monica Gandhi, New Yorker by Dhruv Khullar, etc, and official signs from my govt (all emphasize hand-washing) 1/ 2021-12-14 00:18:57 RT @aaronsnoswell: The first-ever #NeurIPS2021 workshop on the Political Economy of Reinforcement Learning Systems starts in just under 12h… 2021-12-13 22:57:08 RT @hypervisible: “Is anyone surprised” and “this isn’t new” comments on here are rock solid evidence that you have absolutely nothing usef… 2021-12-12 19:34:23 RT @drajm: “we want to reiterate just how important it is for people not to attend social functions if they have any symptoms, even if they… 2021-12-12 07:13:03 I expanded this thread into an essay on how medicine is inherently political 14/ https://t.co/fzeKeHNCaU 2021-12-11 22:54:19 It was an honor to be part of this great group of women data scientists and I'm disappointed about the proposal rejection https://t.co/a6OzUtduK5 2021-12-11 22:51:18 @noam_ker These are steps that would help: https://t.co/zhSPLYISOe 2021-12-11 21:17:09 @_alialkhatib @KashPrime Absolutely. Those with power keep proposing "personal responsibility" for society-wide collective problems https://t.co/hN3luOkEIt 2021-12-11 21:08:55 @KashPrime All signs in the first picture (the collage) are signs I saw in person RECENTLY. None mentioned masks. The 2nd picture from @KashPrime I've only seen on twitter, not out in the wild. 2021-12-11 21:05:57 the public health messaging we are getting vs. the public health messaging we need #COVIDisAirborne https://t.co/MOSe3owHVI 2021-12-11 20:17:24 Here is an example of a more accurate sign #COVIDisAirborne 5/ https://t.co/nXv2XzqJeX 2021-12-11 20:13:40 My state is reopening borders tomorrow with NO masks. There are vaccine passports & Vaccine passports don't account for breakthroughs, waning immunity, kids unvaxxed, & 2021-12-11 20:13:39 All the public health messaging I see is incorrectly based on droplet transmission (wash hands, keep 1.5m distance, clean surfaces, cover cough) with NO mention of masks, ventilation, air filters, #COVIDisAirborne 1/ https://t.co/JKB3Khy9Y6 https://t.co/6LvU0IpiVb 2021-12-11 06:38:10 Reminder that initially "mild" covid cases can still lead to long-term immune, vascular, & Long Covid is a lagging lagging indicator. It will be months before we know what % of Omicron cases result in #LongCovid https://t.co/a8lnlPe1Nm 2021-12-08 02:19:56 RT @ACEMathStats: Interested in the issue of "Responsible #DataScience"? We had well over 100 people at our workshop with @QUTDataScience l… 2021-12-08 01:04:15 @kachangnut Thank you! 2021-12-07 09:33:37 RT @math_rachel: Expertise in East Asia has been ignored. Official public health advice in Japan in March 2020 was better than official adv… 2021-12-07 03:56:57 Our medical & - Disabled & - People from other domains (eg aerosol engineers & - Scientists in Africa & 2021-12-07 03:56:56 Racism is one reason western countries are unwilling to learn from other countries, particularly those in Asia & https://t.co/486T0i5i5d 2021-12-07 03:56:55 Expertise in East Asia has been ignored. Official public health advice in Japan in March 2020 was better than official advice being given NOW in the USA, UK, & https://t.co/YpnABrKjkS 2021-12-07 03:56:54 Disabled & https://t.co/87JkYMUaTx 2021-12-07 03:56:53 There is widespread, well-documented ableism, racism, & Pointing this out does not make you anti-science (I love science, but this is a huge problem). 1/ 2021-12-07 03:13:44 @charleemoyle @timothyjgraham @QUTDataScience Congrats to you both!!! 2021-12-07 00:37:35 @Rainbow_Bodies Yes, many doctors have a dismal understanding of probability, and this creates so many problems. 2021-12-07 00:34:22 @TMS_IEEE I'm glad you found an accurate diagnosis & 2021-12-07 00:31:40 @entoria350 @exceedhergrasp1 Yes, absolutely https://t.co/h8uPG5XBGJ 2021-12-07 00:29:18 RT @1CeliaGreen: The “everyone is going to get Covid” defeatist narrative just absolves government of any responsibility. Everyone is only… 2021-12-06 21:12:54 RT @ShannonVallor: Research friends: what are your favorite AI Ethics articles/chapters written since 2020? Trying to see which gems or und… 2021-12-06 21:10:04 @Abebab Awesome, congrats!!! 2021-12-06 07:30:12 RT @DavidSteadson: Indeed. Some things the literature told us prior to 2020: * coronavirus immunity wanes * sars-cov-1 was airborne * sars… 2021-12-06 04:56:12 If you think ONLY MDs have anything of value to say about health & Medicine systematically excludes most ill & 2021-12-05 22:55:34 @geek_ed_ I am not trashing science. Science supports (and has long supported) that covid is airborne and that post viral illness can be debilitating. The fact that disagreeing with certain prominent scientists & 2021-12-05 21:31:16 @Jimnelson806 Please read this thread: https://t.co/wxSDnIpsp5 2021-12-05 21:29:39 @MarinaArthur12 Please read this thread: https://t.co/Y2qZDaLSB3 2021-12-05 21:27:30 Also, all of this is STILL happening. Last week, CDC director recommended *hand-washing* as a key covid mitigation. I truly want our institutions to be trustworthy, but being able to sit back & https://t.co/OandwyviiQ 2021-12-05 21:27:29 From the start, there was scientific evidence to support the benefits of masks and the risks of post-viral illness. Others held these views too, but they were NOT the mainstream narrative of public health leaders in the West nor of doctors quoted in mainstream media. 4/ 2021-12-05 21:27:28 In March 2020, I said young & 2021-12-05 21:27:26 Some folks tell me experts gave the best advice known at the time, that nobody knew, that the evidence changed. I need to share a few receipts. In March 2020, I publicly advocated for ordinary people to wear masks, at a time when CDC & 2021-12-05 21:27:24 We can not fix public health until we reckon w/ how institutions have failed public’s trust (harmful advice, contradictory rules, overconfidence, disbelief of suffering patients). Patronizing "shut up & https://t.co/zT87eVlyod 2021-12-05 04:00:17 @PutAimovigOnNHS This is not good science 2021-12-05 03:58:26 @PutAimovigOnNHS There are scientists/doctors STILL right now advising the government & 2021-12-04 22:13:03 RT @1CeliaGreen: FINALLY! Now they need to be proactive & 2021-12-03 20:54:42 RT @exceedhergrasp1: @ahandvanish Was just talking about this on seeing a "trust scientists over politicians" tweet. Like yes, often, but a… 2021-12-03 20:53:57 For much of this, it was clear at the time that so-called experts were drawing bad conclusions & https://t.co/tyDO438VIh 2021-12-03 20:02:36 So general public needs to invest a fair amount of time (which many do not have) just to know who to trust, what is true, & 2021-12-03 20:02:35 A problem w telling people "just trust the medical experts" is that they still need enough time & 2021-12-03 09:27:44 @ahandvanish @fitterhappierAJ @Daltmann10 Thank you Hannah, this is very helpful! And thank you for the suggestion to look at research on ME immune dysfunction 2021-12-03 09:24:09 Long covid can involve autoimmune issues, including against the immune system itself: https://t.co/PpdlBeljcW 2021-12-03 09:21:22 Important thread, with links, from @ahandvanish on how #LongCovid can impact the *immune system* in particular (Long covid is multi-system: it can involve immune, vascular, metabolic, & 2021-12-03 04:48:55 @mark_riedl @qutlaw @AdmsCentre Yes, the talks were recorded and organizers said they will be posted. 2021-12-03 04:47:01 @Rpshahmemorial1 @rubyslippahs @ahandvanish @fitterhappierAJ @Daltmann10 Thank you! 2021-12-03 02:47:49 How can we sue big tech? Tesla autopilot had 12 crashes (many with emergency vehicles) before lawsuit could be brought Product liability law + Negligence law + Explainable AI as a pragmatic approach @aaronsnoswell #ResponsibleDataScience https://t.co/ecSb1aj1PP 2021-12-03 02:34:16 There are so many great approaches in explainable AI, and NONE of them line up with the law. -- Prof Dan Hunter @qutlaw @AdmsCentre #ResponsibleDataScience https://t.co/T9m67tz0N1 2021-12-03 02:19:33 Data Ethics Canvas recommended by @skdevitt at the #ResponsibleDataScience Symposium https://t.co/Ql2WrrKERr @ODIHQ @ACEMathStats https://t.co/UAFuDL9apz 2021-12-03 01:45:12 This is such fantastic news!!! https://t.co/VO8zGcSljI 2021-12-03 01:36:11 RT @ACEMathStats: Prof Rachel Thomas @math_rachel speaking at our #ResponsibleDataScience Symposium: the people most impacted by ML/algorit… 2021-12-03 00:27:11 @tmiller_unimelb Explanation in Artificial Intelligence: Insights from the Social Sciences @tmiller_unimelb speaking at @ACEMathStats #ResponsibleDataScience Symposium https://t.co/Bfn9AoMOG9 2021-12-03 00:24:48 Human explanations are - contrastive: we don't explain why B, we explain why B rather than A - social: causal explanation is a form of social interaction - selected: we weigh abnormal, intentional & @tmiller_unimelb on what XAI can learn from social sciences 2021-12-03 00:18:22 @oforbes22 @QUTDataScience @ACEMathStats Thank you Owen! 2021-12-03 00:17:39 RT @oforbes22: Awesome start to the @QUTDataScience @ACEMathStats Responsible Data Science Symposium. Really enjoying @math_rachel’s fantas… 2021-12-02 05:34:46 @fitterhappierAJ Thanks! 2021-12-02 05:27:37 @fitterhappierAJ @ahandvanish @Daltmann10 Thank you! Is there a best reference on the autoimmunity against the immune system? 2021-12-02 05:25:14 @TNTesch @ahandvanish @fitterhappierAJ @Daltmann10 @patientled Thank you! 2021-12-02 05:24:55 @ahandvanish @fitterhappierAJ @Daltmann10 Thanks Hannah! 2021-12-02 04:11:56 @_eleanorina @ahandvanish @fitterhappierAJ @Daltmann10 @gezmedinger @evolutionarypsy @dr_kevinlee @MatthewJDalby Thank you for all your recommendations! 2021-12-02 03:04:15 @_eleanorina @ahandvanish @fitterhappierAJ @Daltmann10 @gezmedinger While that's an interesting thread on potential impact/treatment with vaccines, I'm looking more for a description of the underlying problem, in what ways is the immune system malfunctioning & 2021-12-02 02:50:59 Question: have you seen a good explanation for lay people of how #LongCovid impacts the immune system? cc: @ahandvanish @fitterhappierAJ @Daltmann10 2021-12-01 23:01:07 It is impossible for us to "personal responsibility" our way out of this airborne pandemic. This is a collective problem (the quality of our shared air) and we need collective solutions. Societies involve communal spaces: schools, offices, hospitals. We need to make these safe. 2021-12-01 22:55:21 @SassyAngel1974 I think they are trying to be authoritative, but with incorrect messaging I think underlying issue is that CDC won't acknowledge this is a collective problem (quality of our shared indoor air) requiring a collective solution. Instead pushing "individual responsibility" 2021-12-01 20:40:23 The CDC director message is doubly harmful: - omits crucial info that well-fitting masks, air filters, & - undermines trust in CDC & 2021-12-01 20:26:40 I want the CDC to be a trustworthy institution. I'd love to be able to point to CDC guidance & Emphasizing hand-washing and not mentioning N95s, air filters & 2021-12-01 20:06:54 RT @chrmanning: I agree that, in many cases, reviewability of AI systems is a much more achievable and better goal than explainability 2021-12-01 19:42:41 Well said! Those of us accused of being "alarmists" are not calling for panic. We just want people to address the deaths & 2021-12-01 08:12:26 RT @jljcolorado: In light of the fact that #OmicronIsAirborne, because all variants are #airborne, hand washing is not that important for C… 2021-12-01 07:05:30 @jasonkotenko Thank you! I have also struggled with how to express my thoughts without coming across as anti-science. Here's a related thread: https://t.co/8dc8szpO8S 2021-12-01 04:43:36 22 months into the pandemic, and public health messaging in the West is still too often outdated, inadequate, & 2021-12-01 04:43:35 We've watched some doctors feed anti-vaxxer doubts, at scale, downplaying the risks of covid in children 12/ https://t.co/0ggrsj39qE 2021-12-01 04:43:34 We've had absurd policies (you can gather with friends in a restaurant but not your own home), harmful ones (closing beaches & https://t.co/pFjQPIr7ld 2021-12-01 04:43:33 The experiences & https://t.co/Id6fExoPgT 2021-12-01 04:43:32 People distrust the institutions that don't trust them (@karlitaliliana referring to the disbelief & https://t.co/hPnXrifIFV 2021-12-01 04:43:31 Some leaders lied about the efficacy of masks rather than admit there was a shortage. Which of course backfired when they later needed to ask everyone to wear masks. 4/ https://t.co/JUY7OOC83C 2021-12-01 04:43:30 There has been more focus on whether the public trusts institutions (governments, medicine, public health orgs) than on how those institutions could better earn our trust. 1/ https://t.co/IzQ4EUQdMp 2021-12-01 03:22:54 Yet another example of facial recognition failing: after taking a high-stakes exam, the student is not even notified that she wasn't "recognized." Her results are unavailable & 2021-11-30 23:12:05 Above quotes are from "Reviewable Automated Decision Making", by @jennifercobbe and Jatinder Singh 4/ https://t.co/Os25Vl48zs 2021-11-30 23:11:01 Judicial review of public sector decision-making does not simply assess the decision itself, but the decision-making process as a whole An understanding of human decision-making as a process that begins before the decision and that has consequences that resonate afterwards 3/ https://t.co/rDlhrXbHii 2021-11-30 23:08:34 Explanations focused on how a model has arrived at an output may miss much of what is important. A more holistic view could include information on testing & 2021-11-30 23:05:08 On *reviewability* of automated decision making (ADM), rather than *explainability* Reviewability does not necessarily involve explanations. It is about exposing the decision-making process, including human processes, structures, & https://t.co/Os25Vl48zs https://t.co/Y2mczVovIj 2021-11-30 22:33:42 RT @kprather88: So frustrating hearing scientists say the virus is just going to continue to evolve…like we are helpless. The virus can’t e… 2021-11-30 22:32:08 RT @cfiesler: I've seen these postings too. Though I recognize that people who deeply understand both the social & 2021-11-30 19:53:04 RT @timnitGebru: My gist at this hearing was that the power of these multinational companies needs to be curbed & 2021-11-30 19:34:23 @ChioneNgozi Thanks, I hope you and your loved ones stay safe! 2021-11-30 00:42:02 @ChioneNgozi @kprather88 @denise_dewald @CDCgov @WHO Current messaging in Australia is similarly bad. No mention of masks, ventilation, or air filters: https://t.co/rk8yztilYR 2021-11-30 00:18:17 RT @IEthics: And the full (free) compendium of related materials: https://t.co/NwOtDwWj6I 2021-11-29 09:17:35 "The policy proposal largely deals with issues of defamation, not trolling & https://t.co/sVvHzXzacX 2021-11-29 02:25:43 RT @QUTDataScience: Recently, @math_rachel participated in a Q& 2021-11-29 02:07:33 Also, consider practical processes you can implement to make ethical reflection, deliberation, & https://t.co/lPqJVQ0GGc 2021-11-29 02:07:32 It is crucial to recognize power differences among various stakeholders 6/ https://t.co/xcw3TdsqAZ 2021-11-29 02:07:31 However, there are pitfalls to avoid when approaching "data for good." As @sarahookr has pointed out, it is often an imprecise catchphrase 4/ https://t.co/fha0EiXYR5 https://t.co/1YmUf8ZRKF 2021-11-29 02:07:30 Recent projects in which @QUTDataScience partnered with @CCQld & https://t.co/ZLVfDl2YF6 2021-11-29 02:07:29 My newest post: Doing Data Science for Social Good, Responsibly 1/ https://t.co/HmoJuzzyi8 2021-11-28 09:30:41 RT @antmandan: This latest dead cat by Morrison and Cash on online anonymity is dangerous and stupid. Here is an old thread that outlines s… 2021-11-28 09:14:00 @manku_timma1 There have already been many research studies showing covid can damage the pancreas & 2021-11-28 08:39:39 Thread with more details about the NeurIPS review: https://t.co/2CjogiPRc8 2021-11-28 08:37:28 "The Values Encoded in ML Research" is a valuable & (@Abebab has recently shared the inappropriately rude & 2021-11-28 08:15:37 @LeonDerczynski @qldhealth @QldGov @YvetteDAth @KashPrime Yeah, the opening windows & 2021-11-28 07:49:50 Another important reason for a multi-layered approach: Children under 12 aren't eligible for vaccines yet in Australia. See this thread for more on how covid can impact kids https://t.co/vGrVc0TyQt 2021-11-28 07:45:04 I'm double vaxxed & https://t.co/7CmYwLxfaW 2021-11-28 07:27:31 Here's a great example of a sign that makes clear how we can fight covid @qldhealth @QldGov @YvetteDAth: please consider updating messaging to help keep Queenslanders safe! (sign from Dr. @KashPrime) https://t.co/JKB3Khy9Y6 https://t.co/hcm0N7efCV 2021-11-28 07:01:51 @Choose_Infinity Thank you! 2021-11-28 05:23:17 "In NSW Parliament, the ventilation system has been upgraded to ensure 8 exchanges of fresh air every hour on the advice of health experts. Our children should be afforded the same protection." Just signed (for Australian residents) https://t.co/YeudBdt3oH 2021-11-28 03:35:46 I like this graphic on 5 ways to fight an airborne virus https://t.co/xL3wNMmBxh 2021-11-28 03:20:50 @ongole_pavan Many governments seem to be thinking so short-term, that I'd be surprised if they are even calculating the costs of needed interventions or of mass disability & 2021-11-28 02:20:42 A society that cares more about declining trust in institutions than what institutions have done to deserve trust – and which devotes far more effort towards managing the behavioral psychology of risk than actually reducing risk – is engaged in narrative-making above all else. https://t.co/GkZgrX8uo3 2021-11-28 02:20:40 Failure of legacy institutions to respond appropriately to the pandemic, from March 2020 @aelkus, h/t @RSButner https://t.co/DaPdPZGRcB 2021-11-28 02:20:39 Flaws of countering disinfo w/ appeal to authority: "Worrying about whether we trust institutions without asking if these institutions deserve trust... A program of infantalization – trust that the adults know what is right – will provoke equally infantile resistance." @Aelkus https://t.co/RBdlJQ5BLL 2021-11-27 23:18:24 @kristen_hardy Also: "Victoria Police's secret intelligence unit has infiltrated Melbourne's activist and community groups for two years to gather information on protests against the Iraq War, Japanese whaling and a weapons exhibition." https://t.co/d3LhpcTvNP 2021-11-27 23:16:27 @kristen_hardy Thank you for sharing this Kristen! "Surveillance information was kept in a secret electronic database that had files on more than 1200 Victorians ranging from gay activists to politicians, trade unionists, students, ethnic community leaders and environmentalists." 2021-11-27 22:35:27 Question: does anyone have articles/examples on use of surveillance in Australia, historical or present? Particularly re: surveiling people of color, labor activists, Muslims, immigrants, or other marginalized groups https://t.co/NA5NJr1MLB 2021-11-27 06:34:53 "By 2020, social media had effectively disrupted democratic elections in most modern democracies, and had fundamentally unraveled & I recommend reading in full: https://t.co/HCTBydAjvp 2021-11-27 06:34:52 By the 90s, to even discuss the legacy of racism was allegedly part of the very problem of racism itself... a new post-civil rights-era backlash funded & 2021-11-27 06:34:51 "A powerful & https://t.co/HCTBydAjvp 2021-11-27 06:15:59 I want to re-share this powerful paper by @alvarombedoya, who has been nominated to serve on the FTC around issues of privacy & https://t.co/KrAyEqAhnx https://t.co/NA5NJr1MLB 2021-11-27 05:06:58 @Jeremy11223344 Yes! And legal protections so that employees and students can't be forced to go inside buildings with unsafe air quality. 2021-11-27 01:52:42 The ineffectiveness & https://t.co/yhTM6cofML 2021-11-27 01:48:00 We need governments to: - fund ventilation upgrades for offices & - give everyone N95s - set indoor air quality standards - pay people to stay home if sick or quarantining - educate public Airborne spread - educate public LongCovid & - biomed research #TreatLongCovid 2021-11-26 22:33:16 RT @uche_blackstock: The quickness with which European countries have already instituted travel bans against South Africa, but yet have don… 2021-11-26 09:21:49 @suekeay @Atse Congrats Sue!! That's great! 2021-11-26 02:42:24 "AI inevitably becomes entangled in the ethical & https://t.co/kHx3ShcIMC 2021-11-26 02:42:23 "The truly difficult part of ethics—actually translating theories, concepts & 2021-11-26 02:42:22 Compared to ethics principles in medicine, AI ethics principles lack: 1. common aims & 2. professional history & 3. proven methods to translate principles into practice 4. robust legal & https://t.co/kHx3ShcIMC 2021-11-25 23:32:19 @ShannonVallor Congrats!!! 2021-11-25 22:50:56 RT @ahandvanish: It's time to turn the focus from #LongCovid symptom studies and start studying what biomedical treatments can help. #Treat… 2021-11-25 07:26:12 @jeanburgess Congratulations!! 2021-11-25 06:36:11 This https://t.co/hPnXrifIFV 2021-11-25 05:16:37 "Vaccines reduce risk of Long COVID by lowering chances of contracting COVID in the first place. But for those who do get a breakthrough infection, studies suggest that vaccination might only halve the risk of Long COVID — or have no effect on it at all" https://t.co/G9dts6cZX5 2021-11-25 05:08:19 "When you go into a movie theatre in Japan, the CO2 reading in the theatre is projected onto the screen" "If you're renting a hall for a choir, taekwondo practice, or your child's 21st, you should be able to see a sensor of how fresh air is being controlled & 2021-11-25 05:08:18 Many governments monitor OUTDOOR air & https://t.co/Ce0s8phPFS #COVIDisAirborne https://t.co/lAGaZTHyKM 2021-11-25 01:44:33 @VandekerckhoveJ Why? I get taking your mask off for the dental exam itself, but a waiting room full of unmasked people isn't a good idea (it's fine while still at #ZeroCovid) 2/3 of covid cases don't involve fever, and 50% of covid spread is pre-symptomatic. Temp checks are deeply misleading 2021-11-25 01:23:38 This is partly because of how successful Qld has been thus far in keeping cases & However, now that masks are no longer required & 2021-11-25 01:23:37 - my dentist requires that we use hand sanitizer & - my local hospital has signs about hand-washing stopping covid, but no mention of masks - I regularly meet ppl who don't know about LongCovid 1/ 2021-11-25 00:59:22 To combat Google’s dominance, @acccgovau has considered mandating a ‘choice screen’ for Android to give users option of choosing default search engine when setting up... ACCC also considered putting a stop to the bundling of search engines with devices https://t.co/W4UeRJCPNZ 2021-11-25 00:48:48 As @antmandan noted, regulation in Australia doesn't just impact Australia https://t.co/tDwkb4rERs https://t.co/3MouYhIgtl 2021-11-25 00:48:47 Big Tech's control of resources & https://t.co/NmWsSWKyzQ 2021-11-25 00:48:46 Research institutions & https://t.co/Y6XHQu49xD 4/ https://t.co/E2P6n5F0KK 2021-11-25 00:48:45 Australia's competition regulator found: - Google engages in anti-competitive behavior in digital advertising, which harms consumers & https://t.co/j75Lp7bqXQ - monopoly on search has "stifled innovation and reduced consumer choice" https://t.co/W4UeRJCPNZ 2/ 2021-11-25 00:48:44 on Google, Tech Regulation, & "Google’s investment in Australia’s tech future — a boon for universities or a Faustian bargain?" by @jordanguiao 1/ https://t.co/Y6XHQu49xD 2021-11-25 00:09:44 Responsible Data Science Symposium, Friday Dec 3 in-person at @QUT or online. Cross-disciplinary: experts in statistics, computer science, ethics, law, media, & Register here: https://t.co/z5seRdbdKK @ACEMathStats @QUTDataScience @ABSStats https://t.co/E0i8dkq4gs 2021-11-24 04:26:20 RT @tzimmer_history: I want to be precise: The point is not that there aren’t specific aspects of American politics, society, and culture t… 2021-11-23 05:55:30 RT @TheAngryEpi: As someone who has to take multiple immunosuppressive meds, when I see you with a mask on, I feel loved and supported. It… 2021-11-23 05:23:19 @bag_of_ideas @QldGov @qldhealth I meant ventilation. While vaccination is very important, it has been widely covered here. More on the importance of good *ventilation* in fighting covid: https://t.co/mt9Ex5l9oX 2021-11-23 00:28:23 RT @TarletonG: We can criticize algorithms for being biased, and moderation policies for failing marginalized users. But this will never ch… 2021-11-23 00:05:57 @RobertERitz @QldGov @qldhealth I don't think it's that deliberate a strategy-- I think these graphics were likely created early on during the pandemic and nobody has bothered to update them, perhaps combined with a sense that hand-washing is seen as less controversial/less invasive than mask-wearing. 2021-11-22 22:37:52 RT @sameo416: @dnepo @hughes_eilir @LawtonTri Answer. Because engineers understand physics and can model reality WE DON’T USE RCTs. I’m ge… 2021-11-22 07:13:31 @RobertERitz @QldGov @qldhealth It is best to be honest with the public, as dishonesty contributes to misunderstanding (ppl think they're safe b/c they wash hands often), erodes trust (often ppl can tell when they're being lied to), & w/ more details: https://t.co/7Dq5pwFkdG 2021-11-22 05:19:56 "Interventions by authorities can backfire if they fuel mistrust or treat the public as an adversary rather than people who will step up if treated with respect." Re-reading why lying to public that masks don't work was a bad strategy, @zeynep March 2020 https://t.co/JUY7OOC83C 2021-11-22 03:34:43 RT @exceedhergrasp1: The chronic illness community saw it coming and did our best to prepare folks. #MEAction had a coordinated response. B… 2021-11-22 02:48:37 RT @AugersMonster: This is so so true. I’m not all for locking everyone down, I just want SOME mitigations to keep kids safer while they wa… 2021-11-22 02:09:57 @bazscott I was initially encouraged by Victorian plans for school safety, but it sounds like they haven't been implemented well, if at all 2021-11-22 02:07:13 At my local hospital today, signs from @QldGov saying that we can stop the spread of covid through hand-washing & @qldhealth @QldGov can you update your signage to reflect that covid is airborne and can spread pre-symptomatic? https://t.co/Kn5rtlrosG 2021-11-22 00:56:50 USA laws that let you meet with extended family in a restaurant but not your in own home were clearly contradictory & https://t.co/yL8S7NoLXk 2021-11-22 00:56:49 Zeynep did a great job covering “beach shaming" Dozens of articles focused on beach goers (outside in the sun is a relatively safe place to be) & https://t.co/Asx7nkA4i3 2021-11-22 00:56:48 Ineffective restrictions, like “deep cleans”, plexiglass barriers, hand sanitizer, and closing parks & https://t.co/2YuSo9vL21 2021-11-22 00:56:47 Many people are still wearing low-quality, poor fitting masks (if wearing masks at all). Homemade cloth masks were supposed to be a stop-gap measure while countries ramped up mask production (article from Jan 2021) 9/ https://t.co/alMejmtLdM 2021-11-22 00:56:46 8 of 9 states in Austria got rid of mask mandates in July → covid is out of control now → they have to lock down Masks are LESS RESTRICTIVE than lockdowns. Mask mandates help prevent lockdowns. 8/ https://t.co/TlzFNqFX6l 2021-11-22 00:56:45 A lot of focus has been placed on “personal responsibility”, not so much on what we should expect from our employers, our governments, and our kids’ schools. Droplet transmission was false, but aligned well with neoliberal & https://t.co/5oOaH1HiU5 2021-11-22 00:56:44 Great article from @trishgreenhalgh highlighted that we didn’t just get bad advice in the UK (& https://t.co/KYwVTc3c0z 2021-11-22 00:56:43 The “only elderly & https://t.co/eH8mYCdL0M 2021-11-22 00:56:42 Many people have a false dichotomy that you are either FOR or AGAINST covid restrictions, with no nuance about the TYPE of restrictions or level of effectiveness, much less that eschewing all restrictions → hospitals collapse & 2021-11-18 08:10:31 @noodlefrenzy Thank you, I'm glad we could help 2021-11-18 08:07:33 I find this advice from the Japanese Prime Minister to be so respectful, the honesty about uncertainty. It's not just that we've repeatedly been given wrong advice in USA & 2021-11-18 08:03:21 on valuing randomized control trials to the exclusion of other types of evidence: "Since randomized trials are not considered appropriate in the science of building design and ventilation, this mindset led policymakers to reject a wealth of wider evidence." 4/ https://t.co/zbaCv9JqFF 2021-11-18 07:58:37 WHO emphasized how to protect yourself, rather than what to expect of employers, schools, or govt. Droplet precautions are under control of individuals, resonate w/ neoliberal & 2021-11-18 07:48:51 Contrast in advice from leaders in Japan vs. UK in early March 2020 Japan: warning about poor ventilation, crowds, & UK (& 2021-11-18 07:48:49 "Policymaking is a contact sport involving competing narratives (about problems, how they arose, & 2021-11-18 02:57:39 RT @ahandvanish: Reading quotes from the people who tried to psychologize AIDS, migraines, & "Non-gay… 2021-11-18 02:31:07 Brisbane AI ethics reading group this Tuesday (Nov 23)! There are still spots to sign up: https://t.co/38BUMCTDbZ 2021-11-18 02:03:35 From practical experience on real-world projects: We often may not have to sacrifice accuracy in order to get fairness BUT we do have to deliberately & #MonashPratoDialogue https://t.co/qq0v7OcyOO 2021-11-18 02:01:17 There are many bias/parity measures: Demographic, Impact, False Positive Rate, False Negative Rate, False Omission Rate Decision tree from @rayidghani for which to choose: - punitive or assistive? - can you intervene with most or only small fraction? - among which groups? https://t.co/ymKsf5ANfi 2021-11-18 01:50:02 The focus should not just be on making the AI model fair but rather on *making the overall system & Bias & -- @rayidghani for @MonashDFI lecture on Doing Good with AI https://t.co/ILf1ShRnGK 2021-11-18 01:26:11 Related: https://t.co/H7tLiHcvop 2021-11-17 23:18:40 Many policy makers & - risk of breakthroughs - LongCovid from breakthroughs - risk of reinfection - LongCovid from reinfections - LongCovid in kids - immunocompromised people - ventilation & - that mass disability is bad for economy 2021-11-17 22:56:38 RT @jordanguiao: I wonder why Google is splashing out money in Australia rn - currently live: 1) Inquiry on anti-competitive behaviour in… 2021-11-17 22:25:29 "Regulation in Australia doesn't just impact Australia 2021-11-17 22:22:43 "Google's investment in Australia is like Facebook trying to rebrand as Meta: big tech is trying to reframe the conversation." @jordanguiao 2/ 2021-11-17 22:21:10 Google has been feeling the heat of Australia's regulatory interventions: ACCC inquiry found Google's dominance in advertising is anti-competitive. Inquiry into monopoly on online search. Review of privacy laws, considering restrictions on data collection. @jordanguiao 1/ https://t.co/cyEHu004hR 2021-11-17 21:53:14 RT @holden: @hypervisible But make the first half normal, and the second half you talk to sociologists and security experts. Pros: Good "… 2021-11-17 06:21:16 @mer__edith Great article! 2021-11-17 06:21:02 Tech companies co-opt critique by: - funding & - frame issues of power & - push technocratic remedies - cast elite engineers as arbiters of bias & @mer__edith https://t.co/ey0G1vPSa2 2021-11-16 21:48:50 "Why would a conflicted gov body populated by tech execs recommend 'democratizing' access to infrastructures at the heart of their concentrated power? Because this won't actually reduce that power... they make the water in which AI research swims" https://t.co/8ggHFJnH9x https://t.co/cTlruPr0Vb 2021-11-16 21:32:04 "Anyone who has lost a loved one due to infrastructural failure—whether it is a poorly managed pandemic or an unnecessary war—knows that what they feel is more than personal. Grief, induced by global fascism, is political." https://t.co/XAYJd2J4cR 2021-11-16 21:30:26 "The rhythm of information is perpetually outpaced by the intensity of suffering. The god-like view from above afforded by satellite imagery makes visible the mass graves or incinerated houses but not the people... It's hard to grieve when the algorithm prioritizes novelty." https://t.co/nXTQXhzoLZ 2021-11-16 04:39:14 @FrankPasquale Thanks for writing it! I'm glad you are working on this important topic. 2021-11-16 03:57:33 The medical system is already extremely adept at validating various kinds of black-box systems, as many drugs & 2021-11-16 03:57:32 All of these examples reveal another major challenge: explanations have no performance guarantees. Indeed, the performance of explanations is rarely tested at all. 4/ https://t.co/9XlqiBqNrw 2021-11-16 03:57:31 Increased transparency can hamper users’ ability to detect sizable model errors and correct for them, "seemingly due to information overload." 3/ https://t.co/DVKlgZXrgZ 2021-11-16 03:57:30 Explainability methods of complex AI systems can provide some insight into the decision making process on a global level. However, on an individual level, the explanations we can produce are often confusing or even misleading. @MarzyehGhassemi @DrLaurenOR @AndrewLBeam 2/ https://t.co/OsOICNImXj 2021-11-16 03:57:28 The false hope of current approaches to explainable AI in health care: current explainability approaches can produce broad descriptions of how an AI system works in general, but for individual decisions, the explanations are unreliable or superficial 1/ https://t.co/MXMgA9dMY8 https://t.co/FLmE3xrumL 2021-11-16 02:26:30 @red_abebe @arulebak @Abebab @sendgoodcheers @Geobaido @SekouLRemy Great paper, thank you! 2021-11-15 06:24:48 The more common affective computing becomes, the more these systems will incentivize us to adjust our outward “emotional” states to get computers to behave the way we’d like. @FrankPasquale 3/ https://t.co/nUzMj4offB https://t.co/mrWyQqPqow 2021-11-15 06:24:46 Treating persons as individuals, with complex and evolving emotional lives, is time-consuming and labor-intensive. Attributing some “engagement score” or classifier to them is scalable. 2/ https://t.co/UaHuPf4Fga 2021-11-15 06:24:44 Affective computing is less about capturing existing emotional states than positing them. It defines particular emotional displays as normative and then develops systems for rewarding, imposing, or even policing compliance. @FrankPasquale https://t.co/nUzMj4offB 1/ https://t.co/YxqgEgFSJJ 2021-11-15 02:55:39 The gaze of surveillance tech is always going to wind up focused on Black folks—even if that was not the “intent” of the inventor... Surveillance systems, no matter their origin, will always exist to serve power. -- @hypervisible https://t.co/5rkdvpaIkV https://t.co/jXDE3Y8N1G 2021-11-15 00:09:57 "'Utopian' and 'dystopian' should not be understood as alternatives but co-existing perspectives: What is utopian for Amazon or DoorDash is highly dystopian for their workers." @jathansadowski https://t.co/2sl1EjGHV6 https://t.co/tar9Zw226O 2021-11-14 11:09:02 RT @ToshiAkima: @sidherian @jcooki1 @NjbBari3 @polioandme @RageSheen @OrinCordus @CountryBee2 @takesthecake2 @drbecs @CdrHBiscuitIII @drajm… 2021-11-13 10:37:56 RT @rajiinio: Also - with @AndrewYNg's recent "data-centric AI" push, it can be easy to forget the diverse voices that had spoken up about… 2021-11-13 10:16:07 Above quotes are from "Narratives and Counternarratives on Data Sharing in Africa" by @red_abebe @arulebak @Abebab @sendgoodcheers @Geobaido @SekouLRemy Swathi Sadagopan I recommend reading the full paper: https://t.co/P5jA5R0Gp9 7/ https://t.co/kZ8EN7o609 2021-11-13 10:16:06 Contexts are crucial to understanding data fully 2021-11-13 10:16:04 A common threat is parachute-research, in which non-African researchers benefit from data sharing & 2021-11-13 10:16:02 Power asymmetries are common, resulting in imbalanced authorship & Being on the ground in the field has the largest effect on data, but the data work of fieldworkers is framed as menial labor performed by easily replaceable & 2021-11-13 10:16:00 Dominant narratives around data sharing in Africa often focus on lack, insufficiency, deficit. This framing minimizes the strength, agency, and scientific & 2021-11-13 10:15:59 Stakeholders in the African data sharing ecosystem. Those at the top of the iceberg hold significant power & 2021-11-13 10:15:57 "Who benefits from data sharing in Africa? What barriers exist in the data sharing ecosystem, and for whom? If much of the data sharing practice is shaped by the Global North, how can we ensure that the narrative for Africa is controlled by Africans?" 1/ https://t.co/P5jA5R0Gp9 https://t.co/bohX45xdXU 2021-11-13 06:52:29 RT @hypervisible: This opening graf is so infuriating-- The system deployed by her job doesn't read her face or the hair that grows out of… 2021-11-13 01:54:28 Related: targeted surveillance as a mode of governance (with the case study of unemployment benefits) 12/ https://t.co/hvDmlIcQoK 2021-11-13 01:54:27 These projects often begin with a starting point of artificial scarcity, taking as a premise that we will not provide people with enough food, benefit payments, healthcare, etc 9/ https://t.co/fp5U2TwHm4 2021-11-13 01:54:26 In India, many poor & https://t.co/qeNBcGi5v6 2021-11-13 01:54:25 Read @AmosToh thread for more about automation of social services in Western Europe and @hrw report 5/ https://t.co/mTYRBwBRUI 2021-11-13 01:54:24 Flawed algorithm in UK ignores how often ppl get paid and has led to people going hungry & This is not just a technical error 2021-11-13 01:54:22 In France, updates to an automated system for benefit payments caused errors, delays, & Case workers are unable to correct errors in the system. Some victims coped by *cutting back on food* 2/ https://t.co/55HApxkvTE @hrw https://t.co/59zYvjzn7n 2021-11-13 01:54:21 automation of gov social services (eg food benefits, disability services, unemployment, etc) can be: - implemented with no way to correct errors (software treated as error-free) - smokescreen for policy changes - justify austerity under guise of efficiency - operate at scale 1/ 2021-11-12 23:10:32 I hope will learn from other countries. My state (Qld) is reopening in a month with vaccine passports, but NO masks & I'm still hearing more about hand sanitizer than about fact that #COVIDisAirborne. Not much awareness of #LongCovid either. https://t.co/VQp8zSt2SQ 2021-11-12 09:20:12 @Combsthepoet They are correct! 2021-11-12 09:18:19 @QUTDataScience is inviting more projects from non-profits & For more details & 2021-11-12 09:13:27 I enjoyed this! Thank you to @charleemoyle & https://t.co/n7iPW8Elb5 2021-11-12 09:08:54 @AdmsCentre @jeanburgess @snurb_dot_info @pinkydigital @australian Congrats!! 2021-11-12 04:51:00 Mapping food insecurity in Queensland with an interactive map @fareshare_aus (largest meal charity in ) & 2021-11-12 04:48:47 RT @KerrieMengersen: So cool to see the first fruits of our new @QUTDataScience program on Data Science for Social Good. Thanks to @Farzana… 2021-11-12 04:40:44 This has been the inaugural year for @QUTDataScience Data Science for Social Good, with grad students & @KerrieMengersen & 2021-11-12 04:33:48 At the @QUTDataScience Data Science for Social Good showcase, @oforbes22 sharing about ways to visualize spatial uncertainty for the Cancer Atlas map, using glyphs or hues & 2021-11-12 02:34:08 @pwh67 Thanks Paul, I found the article very insightful (and prescient)! 2021-11-12 00:30:07 Above quotes are from "Targeted! Population Segmentation, Electronic Surveillance and Governing the Unemployed in Australia" by @pwh67 I recommend reading in full: https://t.co/9DpB9d37Jo 5/ https://t.co/ATOaY0yd9F 2021-11-12 00:30:05 Those with ‘risky’ characteristics are subject to ever-greater monitoring, assessment, & 2021-11-12 00:30:03 Increasingly, populations are segmented & 2021-11-12 00:30:01 Classifications are not neutral. The way in which categories are defined and who defines them tell a story of power. 2/ https://t.co/G7SUuUzA2A 2021-11-12 00:29:59 Surveillance is best understood not simply as watching & Surveillance as governance: to what ends is surveillance undertaken, what forms of power operate. 1/ https://t.co/9DpB9d37Jo @pwh67 https://t.co/qL9a4qBo45 2021-11-11 21:43:30 @TheZachMueller @jeremyphoward Congrats on graduating Zach!! 2021-11-11 21:43:16 RT @TheZachMueller: So. Post graduation plans: - @jeremyphoward's Matrix Calculus for DL - @math_rachel's Computational Linear Algebra -… 2021-11-11 21:42:45 @migueldeicaza @ctrl_alt_d @merowing_ @jeremyphoward @GuggerSylvain Aww, thank you! 2021-11-11 21:42:20 RT @migueldeicaza: @merowing_ https://t.co/AGK14708F4 is my favorite. There are others with more math, or more principles, but https://t.c… 2021-11-11 03:53:00 RT @marylgray: I love this paper (and have LOTs of fan feels for the coauthors). They stand on the shoulders of giants in info science & 2021-11-10 05:57:55 "Data makes sense only to the extent that we have frames for making sense of it, and the difference between a productive data analysis and a random-number generator is a narrative account of the meaningfulness of their outputs." 5/ https://t.co/qe9yskyUZJ https://t.co/n9b1yDZ5WQ 2021-11-10 05:57:53 The responsibility that they felt that they could still discharge, and which occupied an ever-greater amount of time, was one towards *organizational processes*, or even to the data itself. 4/ https://t.co/aDjchT6wiB 2021-11-10 05:57:52 The granularity of the data, both spatially and temporally, radically reconfigured the work that they had to do. The very fact of a digital trace produced the necessity of an account, leaving them with less time for their previous responsibilities to parolees and to the public 3/ https://t.co/u6nSkD2o1K 2021-11-10 05:57:50 "Data do not speak for themselves. Data must be narrated—put to work in particular contexts, sunk into narratives that give them shape and meaning, and mobilized as part of broader processes of interpretation and meaning-making." @dourish @Imagenaciones https://t.co/qe9yskyUZJ https://t.co/eqRTEuiOrF 2021-11-09 21:38:03 RT @bennessb: people can be very very very sick without qualifying for any type of support, because in order to qualify for support you nee… 2021-11-09 21:33:51 RT @ahandvanish: Antibody testing to diagnose #LongCOVID disenfranchises women! 1) Males are 4x more likely to retain antibodies 2) 36% of… 2021-11-09 02:43:47 11 Short Videos About AI Ethics: https://t.co/2qgpnBkNY2 An AI Taxonomy: https://t.co/nm07M9C6iY Medicine is political: https://t.co/wZ47xJBzqs Avoiding Data Disasters: https://t.co/kKo4Ba5jOh 2021-11-08 21:36:28 Remember: 80% of eligible is just 64% of the whole population (since kids aren't eligible). We have seen in country after country how trying to rely on vaccines alone & 2021-11-07 07:22:02 RT @UpFromTheCracks: the fundamental violence of predictive analytics is not the algorithm, the algorithm automates enumeration of populati… 2021-11-06 23:20:00 CAFIAC FIX 2021-11-01 19:20:00 CAFIAC FIX 2021-11-01 17:30:00 CAFIAC FIX 2021-08-19 21:18:10 @drjohnflackett Thanks John, it is worrisome! 2021-08-19 07:34:59 Here is a thread on a few ways #LivingWithCovid has impacted the USA medical system, including delayed cancer surgeries, waiting over 10 days for surgery due to gunshot wounds, and medication shortages 7/ https://t.co/4smOQo85uU 2021-08-19 07:34:58 Many who are healthy now have a myth of their own invulnerability-- not realizing that by accident or illness, they could desperately need the medical system, or that a virus could forever disable them. 6/ 2021-08-19 07:34:57 Young, previously healthy people (including children) are getting LongCovid. The narrative of covid only being a risk to the elderly has been false from the start. 4/ 2021-08-19 07:34:56 Living with covid is not just the death count, it is 10-30% of so-called "mild" cases becoming permanently disabled with LongCovid, which can include debilitating neurological effects and constant pain. 3/ https://t.co/GMGutgreXx 2021-08-19 07:34:55 I’m hearing more people in Australia talk about wanting to "live with Covid", even though only 22% of the population is fully vaccinated. #LivingWithCovid (combined with low vaccination rates) means… 1/ 2021-08-19 07:25:56 A man in Houston was shot 6 times while at the grocery store and has been waiting 10 days to have the surgery he needs (but the hospitals are too full with covid patients): https://t.co/W7jLfBWWhi 2021-08-19 07:24:27 San Antonio (city of 1.5 million people) ran out of ambulances last Thursday for 26 mins as coronavirus-related 911 calls hit record levels… "That means we're not transporting heart attacks, traffic accidents or any other sort of medical call." https://t.co/KBYbF62RfI 2021-08-19 03:21:10 RT @mer__edith: This is an impressive achievement! Congratulations to the authors. My critique is structural, not personal. That said, I… 2021-08-19 01:10:24 We find drastic differences in data & 2021-08-19 01:10:22 From goodness-of-fit to goodness-of-data: Goodness-of-fit metrics, such as F1, Accuracy, AUC, do not tell us much about the fidelity and validity aspects of the data. Currently, there are no standardised metrics for characterising the goodness-of-data 7/ https://t.co/8y9kVAAP54 2021-08-19 01:03:29 Missing metadata led practitioners to make assumptions, ultimately leading to costly discarding of datasets or re-collecting data. Lack of metadata & 2021-08-19 01:03:27 ML data collection practices often conflict w/ existing workflows of domain experts. Data creation was added as extraneous work to on-the-ground partners (e.g., nurses, patrollers, farmers) who already had several responsibilities and were not adequately compensated. 5/ https://t.co/uS2m8qafOS 2021-08-19 01:03:25 “In real life, we never see clean data. Courses focus on models & 2021-08-19 01:03:23 Data quality issues in AI are addressed with the wrong tools created for, and fitted to other tech problems—they are approached as a database problem, legal compliance issue, or licensing deal. 3/ https://t.co/OW2flChlWE 2021-08-19 01:03:21 Paradoxically, data is the most under-valued and de-glamorised aspect of AI --Everyone wants to do the model work, not the data work: Data Cascades in High-Stakes AI by Nithya Sambasivan @shivanikapania Hannah Highfill @NaaShomeh @heuristicity @laroyo 2/ https://t.co/BjjOwxZjX7 https://t.co/2YOKheU3aQ 2021-08-19 01:03:19 An overall lack of recognition for the invisible, arduous, & https://t.co/NKSPPmQXQo https://t.co/sbst8nekzi 2021-08-18 21:29:24 RT @albanvillamil: If you taught an ADVANCED "Design for Good" class, what would you cover? I was invited to do a lecture and was underwhel… 2021-08-18 02:42:26 @chris_nafis Thanks for your support back then 2021-08-17 22:54:06 I feel this so much https://t.co/c8Zr3OlxMI 2021-08-17 21:42:05 RT @1Br0wn: “After years of a push to digitise databases in [Afghanistan], and introduce digital identity cards and biometrics for voting,… 2021-08-17 02:39:17 Above quotes from "Co-design and Ethical Artificial Intelligence for Health: Myths and Misconceptions" by @josephdonia @jayshaw29 Great paper, read it here: https://t.co/CnUex3fqfF 2021-08-17 02:39:16 Questions to ask (design humility): - when does co-design substitute other methods of public interest & - what are the limits of design research? - who is sidelined and what might we learn from them? - when shouldn’t we design? 5/ https://t.co/bkzuRAzPyT 2021-08-17 02:39:14 The inclusion of communities in design processes does not necessarily address problems that lead to marginalization in the first place. Indeed, it rarely does. 4/ https://t.co/fHaxdhtxvz 2021-08-17 02:39:11 "Better" involvement does not imply a stronger focus on the whole system (including consequences related to data commodification & 2021-08-17 02:39:09 Novel challenges when co-designing AI: - participation is often unwitting (eg as training data) - AI technologies can be repurposed after deployment - black box nature - hard to account for how data produced by system will be used in future 2/ https://t.co/a8ILi4qKLn 2021-08-17 02:39:07 5 Myths of Co-Design for Ethical ML - ‘Better’ involvement 'better’ design outcomes - Co-design increases agency of patients - Representation reduces risk of harms - Co-design is an inherently ethical approach - All problems can be co-design problems @josephdonia @jayshaw29 1/ https://t.co/1Ua7qxvhno 2021-08-16 21:47:37 11 Short Videos About AI Ethics https://t.co/U8gSlUxdeG 2021-08-16 21:43:26 Getting Specific about AI Risks (an AI Taxonomy) https://t.co/aKIMV2GP8z https://t.co/BWfKcvp2Vq 2021-08-12 22:29:49 RT @cfiesler: @mmitchell_ai I was thinking of having grad students in my ethics class this semester write ethical considerations sections f… 2021-08-11 05:14:58 Diversity is a core value, not a box to check. I appreciate these considerations regarding childcare, visas, CoC, time zones, accessibility, & 2021-08-11 05:14:56 I particularly appreciate this tip on avoiding over-solicited speakers & 2021-08-11 05:14:54 Tips for conference organizers on organizing diverse events, from @JessicaSchrouff @WINRePo1 @WiMLworkshop https://t.co/0dKxbkKAgo 2021-08-02 09:16:48 RT @StanfordAIMI: At the forefront of the discussion about equity in AI, @math_rachel and AIMI co-director @drnigamI will unpack #bias and… 2021-08-02 09:15:48 RT @JanelleCShane: @mark_riedl I think about this quote from @math_rachel a lot. https://t.co/jI9apAD0aV 2021-08-02 09:15:45 RT @mark_riedl: Diagnosing COVID was never a problem that needed to be solved by AI. I’ve not heard any indication that doctors or radiolog… 2021-07-29 03:51:09 RT @timnitGebru: “More than a year later, when Ms. Glasson filed a pregnancy discrimination lawsuit against Google, her counselor told Ms.… 2021-07-28 23:18:34 I'll be at the @QLD_AI_Hub startup drop-in day next Thurs Aug 5, from 12:30-2:30pm if you want to stop by and say hi! https://t.co/PDpFtcmorg 2021-07-28 01:13:43 @kdphd @WHOSTP @AlondraNelson46 @geomblog Congrats!! That is fantastic news! 2021-07-27 05:07:16 The only evidence against Simmons came from ShotSpotter. Initially, the company didn’t detect any gunshots. In response to police, an analyst ruled there had been 4 gunshots (the # fired by police). Police then asked SpotShotter to find a 5th gunshot (supposedly fired by Simmons) https://t.co/0uXGpvID6b 2021-07-27 03:58:22 A ShotSpotter analyst manually overrode the algorithms and “reclassified” the sound as a gunshot. Then, months later & 2021-07-27 03:52:29 RT @EmRoseBaz: Too many people think they're not at risk of #LongCovid because they eat well and exercise. Post-viral illness doesn't work… 2021-07-27 03:32:12 "ShotSpotter analysts frequently modify alerts at the request of police departments—some of which appear to be grasping for evidence that supports their narrative of events." This includes changing type of sound, # of gunshots recorded, & https://t.co/vfpbF2MIEd 2021-07-21 05:58:47 Before we’ve had a chance to think about where we’re going, Copilot has already plotted a path. Not only is this the first piece of info we get (anchoring bias), it’s also an output from an automated decision system (automation bias)– a double-hit of cognitive biases to overcome https://t.co/3NGxSwE9aW 2021-07-20 07:57:50 RT @QLD_AI_Hub: Great turn out for our first @women_in_ai Queensland chapter event with a keynote by @math_rachel and intro by @suekeay! Hu… 2021-07-20 07:57:19 RT @suekeay: “You should get involved in AI”, @math_rachel founder of @fastdotai presents a keynote to the first @women_in_ai Queensland ev… 2021-07-20 07:46:11 .@suekeay kicking off @women_in_ai Brisbane chapter @QLD_AI_Hub https://t.co/c0shyQzpq0 2021-07-17 04:57:43 I'm looking forward to this! ICYMI, here is my talk from the last @StanfordAIMI symposium, the most meaningful talk I gave in 2020: https://t.co/7zN08tKCyf https://t.co/g1HnjnQUba https://t.co/iIxnn4FpCg 2021-07-14 07:24:37 I'm giving 2 talks here in Brisbane this week! July 15 (tomorrow) at the UQ St Lucia computer science seminar @UQ_News: https://t.co/vK8jI9YkCB July 20 (Tuesday) at @QLD_AI_Hub for launch of @women_in_ai Brisbane chapter: https://t.co/ldtVR4oQWY 2021-07-08 01:33:40 @setlinger @altimetergroup @Microsoft Congrats Susan!!! 2021-05-23 06:44:17 RT @mathcolorstrees: Got a story for you. Years ago when I was a new PhD grad working at Twitter, there was a director of product who had s… 2021-05-20 22:28:34 Analysis from @gchaslot on Google search autocomplete for "civil war is " in the month leading up to Jan 6 USA capitol invasion. His findings: Google autocomplete results can be completely uncorrelated to search volumes https://t.co/vLrZO4OIip https://t.co/xHwgidgpZ2 2021-05-20 04:09:57 “It underscores the importance of having people who know the lived experiences in these areas. Just because some parts are part of an algorithm that doesn’t mean everything has to be part of an algorithm.” -- @snowjake https://t.co/KV05iQo9Lm 2021-05-18 01:40:00 RT @jovialjoy: @60Minutes producers spoke to me for many hours. I even spent additional time building a custom demo for @andersoncooper and… 2021-05-18 01:37:06 RT @mariadearteaga: The dissociation I'm feeling on Twitter is a painful reminder of the US-centrism of academic communities focused on 'ju… 2021-05-17 20:41:04 @geomblog @WHOSTP @AlondraNelson46 @BrownUniversity @BrownCSDept @Brown_DSI @senykam @UtahSoC This is fantastic, congrats!!! 2021-05-17 20:39:26 @TheZachMueller @jeremyphoward Congrats Zach! 2021-05-13 18:08:52 RT @homeowmorphism: When I was in undergrad, I went to math seminars and they made me feel stupid. I didn't understand a single word. But I… 2021-05-12 17:45:12 RT @UpFromTheCracks: So it’s fascinating that @google hired Marian Croaks to replace @timnitGebru on the AI Ethics team stating she’s a Bla… 2021-05-12 04:42:30 RT @_KarenHao: Who are your favorite researchers and/or writers producing essential work on the problematic history of how we define and me… 2021-05-11 22:22:32 The book is already out, available here: https://t.co/1hUNXX8XPN 2021-05-11 22:21:22 @skdevitt @DrDaronAcemoglu @erikbryn @katecrawford @shakir_za @red_abebe @robreich @danielsusskind @abenanav @MollyKinder @wsisaac @DrZimmermann That's part of it. Here's the summary and link to my essay: https://t.co/4OaCSn8o1t 2021-05-11 05:10:45 RT @charnewman: I asked Amazon to be released from a noncompete after being sexually harassed. I was told “the company does not release emp… 2021-05-11 05:09:39 Excited to receive my copy of Redesigning AI today, and honored for my essay to be included with such an all-star crew @DrDaronAcemoglu @erikbryn @katecrawford @shakir_za @red_abebe @robreich @danielsusskind @abenanav @MollyKinder @wsisaac @DrZimmermann https://t.co/E8tCboEGuj 2021-05-10 22:31:57 RT @_alialkhatib: hi! i recorded a talk about my paper at #chi2021 - "To Live in Their Utopia: Why Algorithmic Systems Create Absurd Outcom… 2021-05-07 04:22:55 RT @jeremyphoward: Over the last 20+ years I've learned a few things about creating tech startups - here's my attempt to summarize those le… 2021-05-06 22:19:43 @starkcoffee @jeremyphoward @QLD_AI_Hub It was good to meet you! 2021-05-06 22:18:30 @ledell @QLD_AI_Hub @suekeay @jeremyphoward Yes!! 2021-05-06 22:17:43 RT @ledell: @math_rachel @QLD_AI_Hub @suekeay @jeremyphoward I'm struck by the all-women team leading @QLD_AI_Hub... so cool! https://t… 2021-05-06 22:17:15 Great work by @snowjake and @MaxShinnerl. I'm proud that @DataInstituteSF Center for Applied Data Ethics could partner with @ACLU_NorCal on this project 2021-05-06 22:13:51 Important analysis on vaccine equity in CA, and how modelling choices can have a huge impact. Using ZIP codes rather than census tracts means > 2021-05-06 10:08:56 Excited about the launch of @QLD_AI_Hub and the upcoming Qld AI Hub fast ai course! @suekeay @jeremyphoward https://t.co/iefQBJtoTF 2021-05-05 23:12:55 An interesting case study on rankings: applying Foucault to analyze how US News rankings system has permeated law schools so thoroughly https://t.co/8IUi4qOVWL by Michael Sauder & 2021-05-05 22:55:02 RT @QLD_AI_Hub: Did you miss this? We got you! Explore the ethical delivery of #AI. Feat. research and practice, this expert panel explored… 2021-05-05 04:03:51 @isaac_flath @Novettasol @fastdotai Congrats Isaac! 2021-05-05 02:35:37 The above quotes are from "Of algorithms, Apps and advice: digital social policy and service delivery" by @pwh67 https://t.co/7WtXI2QzOj https://t.co/ynNaEBbI5Y 2021-05-05 02:35:35 In the case of RoboDebt (algorithm that mistakenly overcalculated debt, with no human oversight or appeals), the algorithm was used to covertly redefine basic operations & It was not just automation, but a change of government policy & 2021-05-05 02:35:34 A key dynamic arising from digital technology in government is differentiating the population into ever smaller segments, which risks reinforcing social divisions & 2021-05-05 02:35:32 Computerization does not result in the same organization "by different means" but changes what the org does Automation is often justified in the name of efficiency, yet it can paradoxically lead to inefficiency: policy & 2021-05-04 15:03:38 Syllabus from @EthanZ course "Fixing Social Media" with readings & 2021-05-04 06:36:32 Data/technology is often used for additional purposes than why it was originally collected/implemented (which is one of the risks of seemingly "benign" uses of surveillance tech) 2/ https://t.co/WQIXfZRcYG 2021-05-04 06:36:31 Moscow's facial-recognition camera network was initially announced as a way to enforce covid quarantine, catch criminals, & https://t.co/RorbLWEjHG https://t.co/M6Cuk1cMsG 2021-04-29 23:03:53 CelebSET as a benchmark should not be considered as a reward to game or a goal to strive for, but a very low bar not to be caught tripping over. The humble goal of the algorithmic audit is to expose blind spots rather than validate performance. 6/ https://t.co/bgniPZcK0W 2021-04-29 23:03:51 Tension btwn representation vs privacy. Efforts to increase representation of marginalized communities can: - lead to tokenism & - compromise privacy - perpetuate marginalization through population monitoring & 2021-04-29 23:03:50 Auditors need to consider the company's procedure/process, not just final performance on a benchmark. Tax audits that evaluate a company's adherence to a compliance process (and not just submitted financial documents) led to better outcomes. 4/ https://t.co/QwOXPhnH9y 2021-04-29 23:03:49 Audits have to be deliberate so as not to normalize tasks that are inherently harmful to certain communities. Gender classification has harmful effects in both incorrect AND correct classification. Promotes stereotypes and excludes trans & 2021-04-29 23:03:47 External algorithmic audits only incentivize companies to address performance disparities on the tasks they were publicly audited for Microsoft & 2021-04-29 23:03:46 Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditing, from AIES 2020, by @rajiinio @timnitGebru @mmitchell_ai @jovialjoy Joonseok Lee @cephaloponderer with some quotes, but I recommend reading the full paper: https://t.co/Z8CMmztRE8 1/ https://t.co/hRil11FBHK 2021-04-29 02:54:14 @DataSciBae @timnitGebru @csdoctorsister @Abebab @rajiinio @mmitchell_ai @jovialjoy @alexhanna @mer__edith @ruha9 @red_abebe @debraruh @MiaD @hannawallach @LatanyaSweeney @katecrawford Thank you 2021-04-29 01:05:58 @brianavecchione @natematias @themarkup @jackbandy @mathbabedotorg @s010n @mrtz @random_walker Thanks Briana! 2021-04-29 01:05:34 @GretchenMarina @rajiinio @AINowInstitute @alfredwkng @themarkup Thank you Gretchen! 2021-04-29 01:05:10 @ndiakopoulos Thank you Nick! 2021-04-29 01:04:35 I agree, Gender Shades set a gold standard: https://t.co/RwXuT7Snhp 2021-04-29 01:03:07 Great mini-thread from @brianavecchione with more resources: https://t.co/9bzRGcnxjq 2021-04-29 01:01:26 An important point https://t.co/TjOTrcuzxx 2021-04-29 00:59:15 @jonathanstray @CHAI_Berkeley Congrats Jonathan! I'm looking forward to your continued work! 2021-04-29 00:56:20 @jonathanstray I agree. Do you know of legislative proposals around this? 2021-04-29 00:54:26 @dhadfieldmenell Thanks Dylan, I will check it out! 2021-04-29 00:51:35 Mini-thread from @ndiakopoulos with more references: https://t.co/VJNMgJgqLQ 2021-04-29 00:50:07 Helpful collection of resources for investigating algorithms in society, including example investigations, method references, and guidance on FOIA, from @ndiakopoulos & https://t.co/VKCce6b96c https://t.co/G270IGFbOO 2021-04-29 00:43:14 This is an excellent mini-thread from @GretchenMarina with more recommendations on algorithmic audits: https://t.co/x1mrxCiTjj 2021-04-29 00:39:51 @themarkup and @propublica do fantastic independent, investigative journalism, auditing algorithms of major tech companies. Thank you @LeonYin! https://t.co/5PV52pnj6w 2021-04-28 00:16:12 The algorithmic auditing trap: @mona_sloane warned about the risk of enshrining eugenicist tech in weak regulation & https://t.co/HTlQj6mFJ8 https://t.co/AgjA6kZFZ1 2021-04-28 00:12:44 Algorithmic audits will not produce accountability on their own @AlexCEngler analysis on auditing employment algorithms for discrimination https://t.co/IByDcD2fm4 2021-04-28 00:12:43 In "The Case for Digital Public Infrastructure", @EthanZ proposes building auditable & https://t.co/zBrp0AJB0R https://t.co/1wvKQUTLo4 2021-04-28 00:12:41 Question: what are your favorite articles/papers/essays about the idea of external audits for algorithmic systems? 2021-04-27 00:55:01 RT @fperez_org: Even if safe/vaccinated in the US, those of us w/ties to other countries are not OK. I have lost more people in my circles… 2021-04-25 21:49:07 @suekeay @QLD_AI_Hub @_tiamaree @skdevitt @michael_guihot @pwh67 Thanks Sue! 2021-04-25 21:48:53 RT @skdevitt: Looking forward listening to and learning from @math_rachel. Check out Fast AI’s Practical Data Ethics course https://t.co/BN… 2021-04-25 21:48:47 @skdevitt Thanks Kate! I'm looking forward to listening and learning from you 2021-04-24 21:56:45 RT @mmitchell_ai: I really appreciated this article, but there are a couple things that keep nagging at me (thread)...1/ https://t.co/66zYE… 2021-04-24 21:40:45 I'm looking forward to being a part of this @QLD_AI_Hub virtual panel on the Social Implications of Emerging Technologies together with @_tiamaree @skdevitt @michael_guihot @pwh67, (12pm Tues AEST, 7pm Mon Pacific Time) https://t.co/k0k17Rb5El 2021-04-23 21:56:26 The computer system contained "bugs, errors, and defects", yet defendants were prosecuted, convicted, and sentenced on the basis that the data was correct. https://t.co/icwxBEgLoF 2021-04-23 21:42:39 As many as 900 post office employees in the UK were wrongly prosecuted, convicted, and in some cases jailed, after bad data from an unreliable computer system falsely suggested cash shortfalls. https://t.co/CyENV8Fu2R https://t.co/HhLrpPju6j 2021-04-22 04:28:14 @NeuralBricolage @jeremyphoward 2021-04-22 04:28:01 RT @NeuralBricolage: so happy you liked it to complete the story it was https://t.co/No0I8fQOMi: @math_rachel and @jeremyphoward, who ju… 2021-04-22 03:51:40 @WillingCarol Exactly. 2021-04-22 01:51:24 This is such an important point https://t.co/2KFfz0Wc6N 2021-04-22 01:30:02 @mattlungrenMD Thank you! 2021-04-22 01:29:43 RT @mattlungrenMD: Incredible data ethics resources by leaders like @math_rachel should be *required* for all in the medical AI field: Data… 2021-04-22 01:29:27 Read Timnit's full thread for important context: https://t.co/rcsQaDzt1w 2021-04-22 01:28:24 I read Jeff Dean's most recent comments about how AI ethics researchers need to offer solutions and realized I already wrote a thread on this: https://t.co/6cbxoXhpfn 2021-04-21 19:32:45 RT @merbroussard: This morning, I read the new proposed @EU_Commission AI legislation, "Proposal for a Regulation on a European approach fo… 2021-04-21 06:38:08 *reverse centaur*: system where the machine uses the human for support, not the other way around @doctorow Warehouse automation has transformed workers into adjuncts for robots, w/ Amazon injury rates increasing every year in proportion to automation https://t.co/k6kM6FOnCl https://t.co/90hpxT9trJ 2021-04-07 21:48:02 @paul_rietschka My husband and daughter are Aussie citizens, but I need a visa. 2021-04-07 02:42:14 (touring schools, arranging playdates, month 13 of no in-person childcare), learning to drive on left side & 2021-04-07 02:42:13 There are a TON of logistics that go into an international move (not surprising, but it takes a while to wade through) 1/ https://t.co/jjmDv5yNbb 2021-04-05 21:08:04 RT @UpFromTheCracks: “Fired for Speaking Out Against Anti-Blackness and Ethnic Cleansing: Open Letter to Cornell Tech” I’ve written about… 2021-03-30 05:24:49 Even when it loses money, a broken system that consolidates more power will not be discarded. -- @histoftech https://t.co/KzkpwZR3q7 https://t.co/SyAzJ6ryqo 2021-03-30 05:20:05 Every single thing that "happens online," "virtually," and "autonomously happens offline first and often involves human beings whose labor is kept deliberately invisible. Everything is IRL. Nothing is virtual. -- @tsmullaney https://t.co/KzkpwZR3q7 https://t.co/9lQoN6ginm 2021-03-29 22:31:44 RT @BostonReview: We are thrilled to announce our new book: REDESIGNING AI — coming later this spring! Exploring work, democracy, & 2021-03-26 05:39:42 RT @_alialkhatib: i've been a big fan of @nicki_dell's work for years 2021-03-25 21:15:42 @rajiinio @morganklauss @amironesei I really liked the paper! Great work 2021-03-25 21:08:48 This is great news! Congrats to Ali and USF! https://t.co/0Td7bp4Jc1 2021-03-24 03:08:23 The lack of cross-citation across disciplinary lines (found in analysis of 254 AI ethics syllabi) indicates that findings from one discipline are not respected as adequate evidence to inform researchers from another and that, even if valued, are insufficiently accessible. 6/ https://t.co/kLwHb32tSO 2021-03-24 02:47:19 “You Can’t Sit With Us”: Exclusionary Pedagogy in AI Ethics Education by @rajiinio @morganklauss @amironesei https://t.co/8PI11rk6wI https://t.co/T3tsUAh2AM 2021-03-24 02:44:45 For most modern disasters, responsibility lies somewhere between individual agency and bureaucratic control... In actuality, an ethical challenge would not get resolved by an individual or even a group of technologists but will require an inherently interdisciplinary effort. 4/ https://t.co/eypaVaFNUz 2021-03-24 02:38:45 Less about a single engineer’s efforts to enforce their understanding of diverse representation into the model, and more about a form of participatory design where other stakeholders are actively & 2021-03-24 02:38:44 Incidents of algorithmic misuse, unethical deployments, or harmful bias cannot be addressed by developing moral integrity at an individual level. The current issues are the result of collective failure. 2/ @rajiinio @morganklauss @amironesei https://t.co/2hMHwB1Adh 2021-03-24 02:38:43 To rush CS students through simplified, condensed overviews of ethical understanding & 2021-03-22 23:30:25 RT @timnitGebru: When HR sent me the email with "mindfulness" or whatever resources I responded with a link to a @math_rachel article where… 2021-03-22 23:30:16 RT @purpultat: I wrote this series for folks new to AI and ethics, using @math_rachel's excellent blogging tips... and it's continued relev… 2021-03-21 02:30:54 RT @ruchowdh: Q - what are people reading to better understand user agency and choice wrt technology? Purposely casting a broad net. 2021-03-20 03:01:47 Co-founder of the Clinic to End Tech Abuse @nicki_dell & More info & 2021-03-20 02:34:05 RT @rajiinio: Finally, it's out! It's here! In this paper, me, @morganklauss & 2021-02-27 03:50:09 @ledell @jeremyphoward Thank you Erin!! I will miss you too 2021-02-26 03:46:42 I'm a bit behind (lots of logistics involved in an international move) and I appreciate your patience as I respond! 2021-02-26 03:46:20 Thanks so much to all the Australians who have reached out & In case it's helpful, here is an overview of my background: https://t.co/WE7RqVfnvD 2021-02-25 08:44:27 @kim_weatherall @ellenbroad @nicsuzor Thank you Ellen and Kim! 2021-02-25 05:51:54 @hongsupshin Likewise, I'm so glad you came to the workshop in 2019! I hope your are doing well! 2021-02-25 01:08:19 @histoftech Thank you Mar! 2021-02-25 01:07:21 @migueldeicaza @jeremyphoward Thanks Miguel! 2021-02-25 00:19:32 @tkasasagi Likewise, thank you! 2021-02-25 00:18:46 @katecrawford @jeremyphoward Thanks Kate! 2021-02-25 00:18:11 @DiverseInAI Thank you! Definitely 2021-02-25 00:17:08 @kdnuggets @jeremyphoward Thanks! 2021-02-25 00:13:39 @catehstn Aww, thanks Cate! 2021-02-25 00:11:17 @michael_guihot @chrmanning Thanks for sharing the article, I will check it out! I grew up on the Gulf Coast of Texas in a town with higher humidity & 2021-02-24 08:10:44 @sherrying @jeremyphoward Wow, thank you for the kind words! 2021-02-24 08:10:01 @moodymwlee @jeremyphoward Thanks Sarada! 2021-02-24 08:09:11 @einashaddad @jeremyphoward Thanks Einas! 2021-02-24 07:57:54 @brianchristian Thank you for the kind words! And I'm glad you were able to connect with David, he is fantastic! 2021-02-24 07:56:25 @chrmanning Thank you, although I'm curious why you describe Brisbane as an "interesting"choice? 2021-02-24 06:50:18 RT @jeremyphoward: I'm really excited to say that I'm sending this tweet from Australia - our new home! I learned lots in SF and had a gr… 2021-02-24 06:49:05 @summerscope @jeremyphoward Neat, thank you! 2021-02-24 06:47:55 @setlinger @jeremyphoward Thanks Susan! 2021-02-24 06:47:30 @DrLukeOR @jeremyphoward Thanks Luke! 2021-02-24 06:47:12 @ingridavendano @jeremyphoward Thanks Ingrid! 2021-02-24 06:46:53 @purpultat @jeremyphoward Thanks Nalini, I will miss you too! 2021-02-24 06:46:14 @MarietjeSchaake @jeremyphoward Thank you! 2021-02-24 06:46:01 @MsLods Thanks! 2021-02-24 06:43:27 @FemInt @JohnMashey @DrRimmer @QUT @lindasob @InnovationSarah @feraldata @chriseigeland @EliseInTheWoods @terry_carney Thanks, yes it would be great to chat in a few weeks once we are settled! 2021-02-24 06:40:16 @shaananc @CAIDE_UoM That would be great! 2021-02-24 05:20:06 I’m sad to leave my dream job as Director of the Center for Applied Data Ethics at USF. I will be joining the USF Data Institute Advisory Board to stay involved. 5/ 2021-02-24 05:20:05 The last year in the USA has been horrifying. I’m lucky in so many ways: we were able to isolate pretty strictly as a family of 3 (a privilege that many did not have) & 2021-02-24 05:20:04 Big news: @jeremyphoward & 2021-02-22 23:05:22 This is so exciting!! Congrats to Twitter for making such a great choice in hiring! I can't wait to see what @ruchowdh does in her new role https://t.co/Jm3m5Buh9R 2021-02-21 23:51:53 @drchren Also, Samy contributed the initial development of Torch, among many other achievements: https://t.co/z2zWLtAq0M Although he was Timnit and Meg's manager, he was kept in the dark about both of their firings. Here is his statement about Timnit from Dec: https://t.co/kxYCW6FVb9 2021-02-20 22:04:10 RT @rajiinio: People need to understand that what's happening to @timnitGebru & 2021-02-20 05:00:33 That was fast! The video is already up for Dr. McNealy's talk, "Looking Beyond the Prototype: How 'supposed to' makes way for harm" https://t.co/eO1lSTzNHA @DataInstituteSF https://t.co/QKz626Vqn0 2021-02-20 01:59:31 @UpFromTheCracks I'm sorry that I couldn't make it to the live event. It was such a great talk my tweets don't do it justice 2021-02-20 01:52:24 Recommended books: Writings, Du Bois In the Wake Captivating Technology Automating Inequality Catching a Case Killing the Black Body The Protest Psychosis Getting Wrecked Shattered Bonds A Cultural Interpretation of the Genocide Convention The First Civil Right Misdemeandorland https://t.co/WeAr6ksRk0 2021-02-20 01:47:53 If there was a material benefit from the family regulation system (child welfare system), middle class white people would be seeking it out for their kids. 2021-02-20 01:43:28 Calculating the souls of Black folk: predictive analytics in the child welfare system powerful & 2021-02-20 01:39:08 Thread about a harmful example combining both techno-solutionism & 2021-02-20 00:27:10 @mmitchell_ai I'm so angry on your behalf and sorry that you are being treated this way. 2021-02-19 23:56:59 @timnitGebru True, and that's part of what has been so infuriating about Jeff's statements all along. He has to know that we can't possibly believe what he is saying. @_alialkhatib wrote about it re research approval https://t.co/hMVrh8itiF 2021-02-19 23:39:51 I really enjoyed this talk. Check out the rest of the spring data ethics @DataInstituteSF seminar line-up here: https://t.co/KUQhAIJV37 https://t.co/NfqFZGlhaU 2021-02-19 23:35:01 RT @ahandvanish: OMG just as suspected: COVID antibody tests are more accurate in men. They are only 72% sensitive in women! This has *dir… 2021-02-19 23:31:14 Do we really need to automate everything? Who is benefiting and who is being harmed? Sometimes we don't need automation. 2021-02-19 23:24:15 So often technical systems are aimed at solving symptoms, not the root causes. Policy makers need to attempt solving the root issues first. -- @JasmineMcNealy 2021-02-19 23:21:58 We need to rethink data, not as an individual concern solved by individual consent, but as about relationships & 2021-02-19 23:11:43 Central considerations when designing algorithms: - *Who* is the prototype? - Continuous evaluation & - We need to normalize STOPPING the use of a tool when harm is occurring. We don't need to keep using a tool just because we've already started. @JasmineMcNealy https://t.co/7j08C6V6l7 2021-02-19 23:03:00 21 States Are Now Vetting Unemployment Claims With a ‘Risky’ Facial Recognition System: "legitimate claimants have also been rejected by the company’s machine learning & https://t.co/zsMtcLad40 2021-02-19 23:01:14 Is your machine learning solution creating more problems than it solves? @JasmineMcNealy on how in focusing narrowly on one problem, we may be missing many others @DataInstituteSF CADE Seminar Series https://t.co/HPmcNPQiKr 2021-02-19 22:53:28 The first step to improving retention is to stop firing the globally recognized experts who are successfully doing the work, and to hold accountable the people who have wrongly fired them (Dean, Kacholia, etc). 2021-02-19 22:51:43 Dr. Mitchell and Dr. Gebru built THE most diverse team in Google AI, and their team's research wins awards & Dr. Mitchell is being fired the same day Jeff Dean announces Google is *supposedly* going to focus more on D& 2021-02-19 21:58:07 Samy Bengio (a director at Google AI w/ 300 reports, who spoke up for Timnit) is being pushed aside the same week Jeff Dean claims they will start tying exec performance to D& In 2020, Samy hired 39% women, compared to just 14% for the rest of Jeff's org https://t.co/ydb4D4P49t 2021-02-19 09:39:38 RT @justkelly_ok: Here's what happened. I used to be the tech lead of platforms. For months and months, I was left out of meetings and conv… 2021-02-19 09:32:41 RT @_alialkhatib: I’m so excited for @JasmineMcNealy's talk for USF's Center for Applied Data Ethics speaker series, with @DataInstituteSF… 2021-02-19 05:13:24 RT @alicetiara: Does anyone have a syllabus/reading list/etc for disinformation outside US context? Especially interested in post/anticolon… 2021-02-18 23:58:19 I admire & 2021-02-15 22:52:55 RT @alexhanna: Not sure how Google really expects to rebuild trust in folks working in ethics when we're still dealing with this. Mind you… 2021-02-15 19:14:37 RT @sarahookr: Yesterday, I ended up in a debate where the position was "algorithmic bias is a data problem". I thought this had already b… 2021-02-11 19:45:06 @timnitGebru Agreed, it is so exclusionary, and unnecessary https://t.co/jp3FzMEb5S 2021-02-11 16:15:37 RT @timnitGebru: Are we here again? Andrew Ng just did this a few years ago and ppl like @math_rachel detailed how exclusionary this is. An… 2021-02-11 16:15:09 RT @ruchowdh: What’s the incident/ story that got you into responsive/ ethical technology? What stories still stay with you? 2021-02-10 22:59:24 RT @dlowd: No, you don’t have to work ridiculous hours to do meaningful work in AI, and it’s harmful to normalize the idea that long hours… 2021-02-09 00:43:58 RT @IfeomaOzoma: Today, @SenatorLeyva intro'd the Silenced No More Act! I’ve been excited about this bill since last year, when @PeterRukin… 2021-02-08 23:38:55 Great article by @hypervisible and @FoxCahn https://t.co/L16kaS44ae 2021-02-08 23:36:00 "It's crucial to remember that to a large degree, the law enforcement failures that led to Jan 6 were not from a lack of data. Many of the insurrectionists organized openly on Facebook & 2021-02-08 23:35:59 "They never mention the inescapable truth that as these surveillance technologies expand, they won’t be primarily aimed at white extremists. No, these tools will be systematically targeted at BIPOC communities, just as surveillance always has been." -- @hypervisible @FoxCahn https://t.co/608eGX2i8o 2021-02-08 20:02:30 RT @JordanBHarrod: Well, this went about as well as I expected. Definitely appreciate the spotlight from @YouTube, but am now updating my c… 2021-02-07 01:17:30 By focusing too much on the *what*, we miss the challenging & 2021-02-07 01:11:04 Tech-for-good conversations are often wholly separate from policy conversations: governance, questions of rights, & ‘Tech for good’ implies that positive intent drives good technology. That’s not true. https://t.co/0TF2PKXHcR https://t.co/XlwvoIIsm6 2021-02-07 00:40:40 RT @jeremyphoward: There's an open house tomorrow for the remote learning program that totally transformed our family's life in a very posi… 2021-02-05 23:02:52 RT @jeremyphoward: Wow. The PyTorch Lightning founder publicly used baseless patent threats to try to stop new competitors from creating op… 2021-02-05 22:55:25 RT @mmitchell_ai: I am concerned about @timnitGebru 's firing from Google and its relationship to sexism and discrimination. I wanted to sh… 2021-02-05 22:16:42 Join us for the USF Center for Applied Data Ethics seminar series this spring! We have a great line-up: @UpFromTheCracks, @JasmineMcNealy @samanthaa_rr @tripsandflips_ @dev7saxena @Cennydd @xrw, organized by @_alialkhatib @DataInstituteSF More info: https://t.co/JnKQkElQTS https://t.co/uPGvpzUKhV 2021-02-05 21:41:38 @jeggers @hmason Thank you Jana 2021-02-05 18:32:44 @_willfalcon @gridai_ Hi Will, a helpful starting point would be if you could correct false claims that you've made about fastai. This may take significant work on your part, but that work would be a way to show you are aligning your actions with your words. 2021-02-05 18:03:36 There were many features that fastai was incorrectly listed as not having (as well as many features fastai had & https://t.co/0m5DoNeYYr https://t.co/rKUlB0ohiv 2021-02-05 18:03:35 One more example: Falcon published a chart full of inaccuracies, said to comment if something was missing, and then ignored comments 21/ https://t.co/zHBWxL1fMk 2021-02-05 18:03:34 It is also important to note that Falcon raised $18 million in VC funding, likely using the “innovations” of Lightning as a key selling point. This isn't just about open source libraries 19/ https://t.co/J4jLuxeved 2021-02-05 18:03:33 In aggregate, I don’t see how this could just be a “misunderstanding.” Falcon repeatedly made false statements about fastai, did not correct them when pointed out, still has not corrected them, & 2021-02-05 14:38:00 RT @F_Kaltheuner: What are the best tech and tech policy newsletters out there? Interested in non-EU /non-US perspective + newsletters that… 2021-02-04 21:50:53 @samuelekpe Thank you! 2021-02-04 21:45:09 RT @choldgraf: One of the *most annoying and tiring things* is when other products/organizations/etc decide to compete with you, and embark… 2021-02-04 21:44:59 RT @cristi_vicas: I was a big fan of TF/Keras. I said that Theano lost bc of larger community around TF. In 2017ish I saw fastai and I was… 2021-02-04 21:44:17 I want to credit @GuggerSylvain for his crucial role in developing fastai. He helped make fastai what it is. Also, thank you to the fastai community, and everyone who invested time & 2021-02-04 20:49:21 There is also a common bias at work here: the false belief that you can’t be doing something to increase diversity/include outsiders AND being doing state-of-the-art work, as though these two are in conflict (they are not) 15/ 2021-02-04 20:47:23 The fastai library has been used in many research papers. Here are two partial lists 14/ https://t.co/m7ovTJE1qF https://t.co/SRmnxXJFo9 2021-02-04 20:45:35 In 2018, fastai was winning competitions with new research & 2021-02-04 20:45:34 This https://t.co/ktYtgBpxGr win was covered by the Verge 11/ https://t.co/5ysvbMp6Xb 2021-02-04 20:40:49 The below statement is false (again, we taught & 2021-02-04 20:34:54 We were often implementing new papers as they came out during the course. One of the goals was for students to learn how to implement new papers on their own. We updated the course & 2021-02-04 20:33:51 I was frustrated by how cliquish deep learning was at the time (I first became interested in it in 2013), and even as a math PhD & https://t.co/Em7TsGzzju https://t.co/gcqJkD150N 2021-02-04 20:29:49 He laid out our goals, and 2 & 2021-02-04 20:29:48 Because our MOOC is so well-known, some assume https://t.co/ktYtgBpxGr is just for beginners, yet we have always worked to take people to the state-of-the-art (including through our software library). 2/ 2021-02-04 20:29:47 The founder of @gridai_ (@_willfalcon) has been sharing some falsehoods about fastai as he promotes the pytorch lighting library. I want to address these & 2021-02-04 03:25:36 RT @mekkaokereke: Now realize that during their teens to 30s, Black men have as much chance of being killed by a cop, as you do of dying of… 2021-02-04 01:51:19 RT @vineshgkannan: Yesterday was my last day at Google. I left because Google's mistreatment of @timnitGebru and @RealAbril crossed a perso… 2021-02-03 21:31:15 RT @UpFromTheCracks: Feb 12 Come thruuuu https://t.co/8keLtph4YZ 2021-02-03 21:29:55 Exciting line-up for @DataInstituteSF data ethics seminars this spring! More details & cc: @craignewmark https://t.co/MuRaMdwEdD 2021-01-25 20:21:09 Lots of great resources on algorithmic bias in the replies: https://t.co/w8o8kgPtah 2021-01-22 18:31:38 Massive algorithmic systems are functionally similar to sprawling administrative states. We need to understand bureaucracy to understand why algorithmic systems can become inaccurate, absurd, & @kharijohnson on @_alialkhatib's research https://t.co/h6hLewDMPw 2021-01-19 19:50:22 Few things I wrote last year: The Problem w/ Metrics is Fundamental Problem for AI https://t.co/bg0elsbBd9 Ethics chapter, fastai book https://t.co/qeTJZOPbg8 Medicine's Machine Learning Problem https://t.co/bhHkbjny0K Far-reaching Impact of Dr. Gebru https://t.co/4i1y4P27kM 2021-01-19 00:15:19 RT @IjeomaOluo: Let's recognize that in Dr. King's time, the majority of white people in America disapproved of his work. Recognize that he… 2021-01-18 19:28:05 @JulienneLaChan1 @WendyNorris In this thread, I share my data ethics syllabus (course has no prereqs and is heavily focused on societal impacts), including the twitter handles for authors of all the assigned reading: https://t.co/91BURlpgWr 2021-01-18 01:49:21 RT @GuggerSylvain: Excluding sex might be an easy way to remove all pages with sexual content, but it will also exclude any page mentioning… 2021-01-16 21:42:56 RT @willie_agnew: The NLP research community bears a lot of responsibility for this: spending millions to train a big version of a standard… 2021-01-16 00:53:39 RT @_alialkhatib: i didn't get to go far on this but "weakened civil society" happens when AI people disqualify the public from engaging wi… 2021-01-16 00:53:10 Also, check out @_alialkhatib's excellent guest lecture, Losing the Forest for the Trees: How we Measure and Lose Track of Things https://t.co/gagf6kMHCc https://t.co/ZttHnULalk 2021-01-16 00:45:32 Apply for @DataInstituteSF Intro to Data Ethics Course, to be taught online Tuesday evenings March 23-April 27 by @_alialkhatib https://t.co/CPYW4YHYE5 2021-01-16 00:43:27 These quotes are from @_alialkhatib #CHI2021 paper "To Live in Their Utopia: Why Algorithmic Systems Create Absurd Outcomes" https://t.co/FmIOCD8S4Y https://t.co/RwKFDbzPdg 2021-01-16 00:41:02 People are obliged to find ways to conform to the patterns found & 2021-01-16 00:30:19 Factors leading to harm & - treating simplified models ("abridged maps") as reality - unscientific optimism of how comprehensive planning will improve life - authoritarian state - weakened civil society @_alialkhatib arguing how "Seeing Like a State" explains AI harms https://t.co/9Ir4XFao0o 2021-01-14 22:05:15 RT @_alialkhatib: hey! i was on @sfcivic earlier talking about twitter & 2021-01-14 17:28:29 Good thread in response to Yann LeCun saying "FB is an easy target for scapegoating, but let's look at the data" (that data is not available) https://t.co/xL4QeyYw46 2021-01-14 17:21:37 @F_Kaltheuner I'm so glad to hear this! 2021-01-13 22:56:29 @AngeBassa Congrats Angela!!! 2021-01-13 19:56:41 "Cloth masks, especially homemade ones, were supposed to be a stopgap measure. Why are so many of us still wearing them?" @zeynep and @jeremyphoward https://t.co/alMejmtLdM https://t.co/q0Bm8vgzXk 2021-01-13 19:06:46 RT @IfeomaOzoma: In 2019 while at @Pinterest, I wrote in an eng focused trust & 2021-01-13 16:47:15 This is part of an ongoing pattern where @JeffDean publicly praises @timnitGebru's work when he thinks it will make him/Google look good, while simultaneously being dehumanizing & 2021-01-13 07:27:20 My mistake, it is 4 papers, not 3. Even more absurd. Jeff Dean also links to: Diversity and Inclusion Metrics in Subset Selection https://t.co/sZTFQ3OGVO 2021-01-13 07:03:18 @JeffDean @GoogleAI The inconsistency between how you treated Dr. Timnit Gebru (firing her & https://t.co/82ySs0qd0N 2021-01-13 06:43:41 Here are the links so you can skip Jeff's post: Model Cards for Model Reporting https://t.co/cjlh0sVkiT Saving Face: Investigating the Ethical Concerns of Facial Recognition Auditing https://t.co/Z8CMmztRE8 Closing the AI Accountability Gap https://t.co/hodbTVtloB 2021-01-13 06:39:35 Jeff actually links to *3* of @timnitGebru's papers in this post (after firing her last month & Model cards, Saving Face, and Closing the AI Accountability Gap https://t.co/huWK5iy9tZ 2021-01-13 01:19:29 "Nearly every Black woman interviewed for this article said they would be hesitant to recommend that a mentee or student apply to Google." https://t.co/PpZS0JVdA4 2021-01-13 01:17:44 Google's pipeline for Black women: "You put a machete at the beginning of your pipeline, you put a machete at the end of your pipeline, what do you think happens in the middle? You completely shattered it" @csdoctorsister on Google firing @timnitGebru & 2021-01-12 22:31:50 I'm honored to be included with such a fantastic group of scholars. Check out @luke_stark's syllabus & 2021-01-12 21:50:09 @DataSciBae Thank you! 2021-01-12 21:49:47 RT @DataSciBae: #30DaysOfAIEthics Day 12: I really love this playlist of short #DataEthics videos from @math_rachel. The first shows why e… 2021-01-12 00:30:25 This comprehensive review on masks was first released April 2020 and has been the #1 most read preprint of all time on https://t.co/yNASRGAxnR. I'm glad to see it was finally published in PNAS. Congrats to the team! https://t.co/BzDhc48v3O 2021-01-11 19:01:11 RT @math_rachel: I wrote a post about @random_walker's AI taxonomy and his insight that different categories pose fundamentally different r… 2021-01-11 17:28:10 RT @rcalo: Don’t let anyone tell you we need new surveillance tools or powers to prevent the next insurrection. The problem has always been… 2021-01-11 17:10:59 RT @UpFromTheCracks: This seems particularly concerning -albeit unsurprising- thinking about public sector data sets ie Child Welfare, Home… 2021-01-10 20:07:56 Journalists & - anonymized browsing history dataset of 3 million Germans - de-identified medical records of 10% Australian population - anonymized NYC taxi trajectories, London bike sharing trips, Riga subway data https://t.co/o4Ev90dJrx 2021-01-10 20:07:55 99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes https://t.co/mbwbtrZAfQ @cynddl @yvesalexandre h/t @snowjake https://t.co/Xt02U7w8zw 2021-01-07 19:32:40 RT @holden: I cannot stress this enough. Until there are consequences this will continue to get worse. If you are in *any* position to en… 2021-01-07 02:02:27 RT @jbouie: it’s not just the storming but the impunity that makes this so reminiscent of late-Reconstruction era attacks on elected offici… 2021-01-06 20:02:48 I wrote a post about @random_walker's AI taxonomy and his insight that different categories pose fundamentally different risks https://t.co/flOz2Oy0DK https://t.co/LZjpPF2P60 2021-01-05 23:58:07 RT @ruchowdh: How can startups implement technology ethically? I'm an advisor for @startupsandsoc - brainchild of @WilneidaNegron and @lyel… 2021-01-05 21:42:46 Good advice on what to do when you feel completely stuck on a research project, from @maithra_raghu's reflection on her machine learning PhD journey https://t.co/2ICOzkcvPn https://t.co/HoS5b8HeRF 2021-01-05 19:46:56 RT @jason_koebler: SCOOP: @LaurenKGurley obtained a spreadsheet that Google's union-busting consultants used to try to prevent a hospital f… 2021-01-05 19:07:44 RT @jeremyphoward: Stunning new work from @math_rachel in @BostonReview: "Medicine's Machine Learning Problem: As Big Data tools reshape he… 2021-01-05 17:04:34 @atg_abhishek @amironesei @_alialkhatib @ruchowdh @usfca @DataInstituteSF @craignewmark Yes! You can subscribe to the CADE newsletter here: https://t.co/do9jQzfqn0 2021-01-05 17:03:00 @OpenSexism @BostonReview That quote is from @rajiinio's article: https://t.co/eXM0l4PTBS 2021-01-05 16:49:19 USF Center for Applied Data Ethics round-up: Medicine’s Machine Learning Problem, A Genealogy of ImageNet, How Google’s Meltdown Could Shape Policy. Articles by or quoting @amironesei @_alialkhatib @ruchowdh https://t.co/935qWT8lnU cc: @usfca @DataInstituteSF @craignewmark https://t.co/wMWDc8ZQT3 2021-01-05 00:36:10 @mona_sloane Thank you Mona! 2021-01-05 00:35:59 RT @mona_sloane: Great piece (-- > 2021-01-05 00:35:36 @luke_stark @BostonReview Thank you Luke, I'm honored to be on your syllabus! 2021-01-05 00:33:18 @Abebab Congrats!!! 2021-01-05 00:32:58 RT @IfeomaOzoma: The Alphabet Workers Union announcement makes me think of dozens of people - mostly women of color/queer - who made this… 2021-01-05 00:32:35 This is such great news! https://t.co/PlSvT0xNmp 2021-01-04 21:52:58 @baxterkb Thank you so much Kathy! 2021-01-04 20:50:24 @JanelleCShane Thank you Janelle! 2021-01-04 20:49:45 I'm not going to link to the "lack of diversity in AI is just a pipeline problem" article, but I wrote a relevant, long thread recently, linking to a ton of research on this (spoiler: it's *not* the pipeline, and it's *not* going to be fixed by just waiting) https://t.co/k3o13l88ro 2021-01-04 20:42:42 @hypervisible @BostonReview Thank you!!! 2021-01-04 19:40:21 RT @JanelleCShane: Essential reading for anyone thinking of using machine learning in healthcare. https://t.co/6xjWc27kKW 2021-01-04 17:28:44 Machine Learning + Medicine: - Data can be incorrect, missing, biased - ML systems can centralize power - Med system is already disempowering & - Expertise of patients often overlooked - Need to focus on power & https://t.co/vLmxa0bdzP 2021-01-04 17:24:55 I wrote about Medicine's Machine Learning Problem for @BostonReview. Please check it out: https://t.co/vLmxa0bdzP 2021-01-02 02:33:58 "There was no question that these algorithms were ever not going to be racist. What would a not-racist predictive policing program look like? You would have to reimagine prediction. You would have to reimagine policing. You would have to reimagine the history of computation..." https://t.co/trB33s2vZj 2021-01-02 01:47:51 @frede_sib @hamidtasnuva Thank you for the rec! 2021-01-02 01:45:29 "The focus on 'surveillance' has a depoliticizing effect on the work we do... These technologies aren’t just creepy. These technologies don’t just make the subject feel watched, or like they can’t express themselves. These are violent technologies—carceral technologies." https://t.co/GJ3VPuHGVp 2021-01-02 01:30:01 @emilymbender @rajiinio Yes!! Those are both fantastic articles 2021-01-02 01:29:07 RT @emilymbender: @math_rachel This piece by @rajiinio was also amazing --- I haven't referred back to it as much, but it's also much more… 2021-01-02 01:29:02 RT @emilymbender: @math_rachel Picking a favorite is hard, since I feel like I have to systematically remember everything I've read, but I… 2020-12-31 22:23:38 RT @kdexd: @math_rachel "Large Image Datasets: A Pyrrhic Win for Computer Vision?" by @vinayprabhu @Abebab https://t.co/CqVHs15Shw Unexpe… 2020-12-31 22:22:19 "This crisis makes clear that the current AI research ecosystem—constrained as it is by corporate influence & 2020-12-31 22:19:22 There are close ties between tech companies and ostensibly independent academic institutions... the two groups look awfully similar. Moreover, the top CS departments accept copious amounts of Big Tech research funding. https://t.co/Px92Miljbp @alexhanna @mer__edith https://t.co/kOAenvwsBs 2020-12-31 20:02:46 @MarcosCarreira @zeynep Yes, that's a great article! 2020-12-31 20:01:38 @mguariglia @culturejedi Yes, that article was great! 2020-12-31 20:01:12 @arthvid Yay! 2020-12-31 20:00:33 RT @amy_tabb: @math_rachel I'm sort of bad at picking a favorite, but I loved this paper and it was a fun read filled with the authors sayi… 2020-12-31 18:06:04 What was your favorite paper/article/essay you read in 2020? 2020-12-30 23:31:43 @ergodicthought @az_jacobs @hannawallach Thanks for the reference. If you haven't already, you may be interested in Ali Alkhatib's "Street-Level Algorithms" (on differences w/ street-level bureaucrats) & https://t.co/ins5z8WDn8 https://t.co/UYNM0ra4Bg 2020-12-30 23:15:35 RT @bigblackjacobin: OK this will be a long ass thread of academic & 2020-12-30 22:45:00 @hannawallach @FAccTConference I look forward to reading it! 2020-12-30 22:06:17 "Measurement and Fairness" by @az_jacobs @hannawallach https://t.co/7iKSBPWSTI https://t.co/9EFodGTXWj 2020-12-30 22:02:49 Some types of Validity content: does measurement capture everything we want convergent: match other measurements predictive: related to other external properties hypothesis: theoretically useful consequential: downstream societal impacts reliability: noise, precision, stability https://t.co/NKCoU4StSu 2020-12-30 21:59:22 A measurement model is "valid" if the theoretical understanding matches the operationalization. There are many ways validity can fail. https://t.co/RBjAcDhK6h 2020-12-30 21:54:01 In computational systems, we are often interested in unobservable theoretical constructs (eg "creditworthiness", "teacher quality", "risk to society"). Many harms are result of a mismatch between the constructs & 2020-12-30 18:11:47 @JulienneLaChan1 @VaroonMathur It's depressing. @Combsthepoet shared some demographic stats related to this: https://t.co/X7TyC7C02h 2020-12-30 15:00:12 RT @VaroonMathur: This is the first story I’m seeing of two independant algorithms working in tandem to incarcerate - facial recognition co… 2020-12-30 05:12:59 RT @rajiinio: "He is the third known Black man to be wrongfully arrested based on face recognition." My question is - how many more does t… 2020-12-29 20:25:57 Here's a thread about this in the context of corporate diversity efforts. It would be great if there was a conflict-free, everyone-feel-good way to address the toxic workplaces driving people out of tech, but there is not: https://t.co/pul2aiOC1N 2020-12-29 20:18:47 The guy who said my tweet is "un-empathetic" & 2020-12-29 19:55:24 It is a myth that there is a form of advocacy that involves only positivity & 2020-12-28 23:59:43 RT @timnitGebru: This chat is how I found out I was resignated. From my direct report in shock who was trying to reach me. I then informed… 2020-12-28 23:41:40 RT @wimlds: @sandhyasimhan @math_rachel That's good news in this individual case and we're happy to hear it. However, it's important to al… 2020-12-28 14:31:13 RT @LofredM: Great content @math_rachel! Thanks for sharing. 2020-12-28 14:31:11 @LofredM Thank you! 2020-12-28 01:07:35 @cgarciae88 If he has changed his views from the past, it would really help for him to announce that: https://t.co/5JXUYPak5q 2020-12-28 01:06:06 @thusinutz @sandhyasimhan https://t.co/7NuCZfJ3iK 2020-12-28 01:05:24 @desertnaut @sandhyasimhan I think the language that he chooses for the newsletter is significant, and I don't know how many people clicked the link. If he agrees she was fired, why not say that? Google has been using the language "departed", which similar to "parted", usually doesn't mean "fired" 2020-12-28 00:46:13 @sandhyasimhan I also don't want to distract from how harmful it is for Andrew to side with Google against Timnit in his newsletter. This thread is useful in addressing the debate of whether Timnit "parted ways" or if she was fired. 3/ https://t.co/jig5ZqyrUA 2020-12-28 00:40:20 @sandhyasimhan For more background on how demanding long hours discriminates against many people with chronic illnesses (not just me), I wrote this 2/ https://t.co/jp3FzMEb5S 2020-12-28 00:36:20 @sandhyasimhan I'm glad to hear that. It would be helpful for him to post if he has updated his beliefs, since many look to him as a role model I'm a mom too, & 2020-12-27 18:58:31 RT @csdoctorsister: @math_rachel Say this louder...back in Nov 2017 he tweeted about diversity in AI, I responded --- received several "lik… 2020-12-27 18:12:03 Andrew Ng: - suggesting Google didn't fire Timnit - has never graduated a woman PhD student - requires employees to work 70-90 hrs per week, which is impossible for many folks w/ chronic illness & 2020-12-27 00:22:10 If you know folks who prefer video, here is a 9 minute video covering some of this material (the need to start at the *opposite* end of the pipeline: the toxic workplace, bias in performance review & https://t.co/OBi3fUqzZP 2020-12-27 00:15:36 Here are a few articles I've written on the topics from this thread (all link to a ton of research) 22/ https://t.co/NGNSxnnAdt https://t.co/1t6T4LKqiA https://t.co/NkVc5Yseo9 https://t.co/WThizQ3uQd https://t.co/chSF29FQ7O 2020-12-26 22:52:30 Advanced Tech is *not* a Substitute for Policy/Regulation Case study on what causes tech companies to act faster: 5 years of warnings about escalating genocide vs. threat of significant fine? (A: the fine) Many AI ethics issues are Human Rights issues https://t.co/ypqILfYiUA 2020-12-26 22:47:40 Steps towards addressing the AI diversity crisis https://t.co/3FULl2axBc 2020-12-26 22:46:41 Tech Ethics Practices to Implement at your Workplace (including work from @scuethics @ShannonVallor @IEthics @megyoung0 @TechPolicyLab @alexfeerst ) https://t.co/4JZ2rRnn7M 2020-12-26 22:43:15 What you need to know about disinformation (as a machine learning practitioner): https://t.co/oMyN5oyZAq 2020-12-26 22:36:12 7 Questions to Ask About Your Machine Learning Project https://t.co/k5QnhKfJQw https://t.co/jnRq0A1AJF 2020-12-26 22:34:16 The idea of bias quickly becomes too generic to be useful. Different types of bias have different sources & https://t.co/0xeCv7JTx1 2020-12-26 19:47:50 The Problem w/ Metrics, Feedback Loops, & Goodhart's Law: when a measure becomes the target, it ceases to be a good measure. Yet this is the underlying premise of many ML products & https://t.co/vlt46xPrEr 2020-12-26 19:43:39 What are Ethics & - 3 Case Studies to know about - Is this really our responsibility? - What is ethics? @scuethics - What do we teach when we teach tech ethics? @cfiesler https://t.co/ryojTAU18G 2020-12-26 19:36:17 I made a playlist of 11 short videos (most are 7-13 mins long) on Ethics in Machine Learning This is from my 2 hrs ethics lecture in Practical Deep Learning for Coders v4. I thought these short videos would be easier to watch, share, or skip around https://t.co/bLHhdPLch4 https://t.co/eQIVwtJ38k 2020-12-26 19:26:56 Multiple women of color at Google reported a manager for harassment & 2020-12-25 04:24:55 RT @timnitGebru: I can’t get over this. When they respond to an article about the person they fired by saying how proud they are of their a… 2020-12-25 04:03:41 RT @kat_heller: @math_rachel @timnitGebru @JeffDean Flagging @greg_corrado and @dtfeinberg here, since they were my management chain at the… 2020-12-24 17:55:53 I agree with @karlitaliliana's thread that we need a new set of laws to make workplace racial discrimination actually illegal. Even though I've been writing about what companies should do, this does not replace the huge need for regulation 21/ https://t.co/72QHAbgG84 2020-12-24 17:20:55 When people expect that diversity can be improved with relatively "easy" gestures (eg: sponsor Grace Hopper Conference, tutor middle school girls, host a women's panel), they aren't prepared for the conflict that arises from anyone advocating for deeper, meaningful change. 19/ 2020-12-24 17:20:54 To be clear: improving diversity in your org will require substantial changes (to your promotions process, performance reviews, hiring, & 2020-12-24 16:29:31 RT @EricaJoy: @math_rachel @timnitGebru @kat_heller @JeffDean my hypothesis: they have deep, unshakable trust in the systems that have serv… 2020-12-24 16:28:47 @EricaJoy @timnitGebru @kat_heller @JeffDean Thank you Erica, that is a helpful insight 2020-12-24 16:10:30 @timnitGebru @kat_heller @JeffDean It's really messed up that Jeff & https://t.co/43uVQhx2Re 2020-12-24 01:00:47 This dynamic happens a lot: a powerful leader (like Jeff Dean) follows the advice of a much less powerful person in HR to do something terrible I know HR is there to protect the company, not you, etc, but I don't fully get why leaders of Jeff's stature don't stand up to HR more https://t.co/fhhTUTxGnr 2020-12-23 23:22:59 RT @timnitGebru: That they thought it was ok to put IN WRITING that they immediately “resignated” me for an email talking about DEI issues… 2020-12-23 21:52:47 @rajiinio @morganklauss @amironesei @FAccTConference Congrats!! I can't wait to read it! 2020-12-23 19:34:55 RT @rajiinio: What? We already have solutions, people have been working on this for so long. We just never implement the solutions. https:/… 2020-12-23 19:28:37 Related: Harassers are often more litigious than their victims, so companies prioritize harassers 17/ https://t.co/wcU3neZLB0 2020-12-23 19:28:36 Another key obstacle to diversity is that HR & https://t.co/3CjMFTPzYo 2020-12-23 17:30:01 RT @baykenney: One solution is to hire more underrepresented minorities. Just hire them. Period. 1/n 2020-12-23 17:18:19 More insight on the interview process: 14/ https://t.co/3nqYwRZBWL 2020-12-23 17:16:42 Also, read these tweets from a former Googler about how the interview process is set up so that one person can easily nix women and people of color. This is true at a lot of tech companies 13/ https://t.co/u3VHJ0O9C1 2020-12-23 17:13:44 Research shows people like to hire people like them. You will need to overhaul your interview process: 12/ https://t.co/WThizQ3uQd https://t.co/ZhuLyFCbHF 2020-12-23 17:11:32 Men are more likely to receive actionable feedback in performance reviews, whereas women receive vague personality criticisms. Again, guess which is useful for getting promoted? 11/ https://t.co/bOBZyXJiWX 2020-12-23 17:11:31 Action item: think of 1 person from an underrepresented group at your workplace, and champion their work: advocate for their ideas, praise them to leadership at your company, or recommend them for an interesting, high-impact work project. 9/ 2020-12-23 17:11:30 When men are mentored, they receive public endorsement of their authority. When women are mentored, they are given advice on how they should change. Guess which of those is actually useful for getting promoted? 8/ https://t.co/NkVc5Yseo9 https://t.co/C72F6xVWTc 2020-12-23 17:11:29 So what works? The #1 thing to do is to make sure that the Black women at your company actually have a fair shot at success. This also means you will have to get rid of the people who are actively sabotaging them 6/ https://t.co/NkVc5Yseo9 https://t.co/SMtukbip9v 2020-12-23 17:11:28 First, what doesn’t work: shallow, showy diversity efforts (even if they are well-intentioned) aren’t just ineffective, they actively cause harm. Spend time thinking through your strategy & https://t.co/1t6T4LKqiA https://t.co/LLxflOMPWF 2020-12-23 17:11:27 Interested in improving diversity in AI, or in tech in general? I have done a bunch of research on this and have some advice 1/ https://t.co/c439bAFirU 2020-12-23 16:19:56 The person who made the "Yann LeCun vs The Mob" video also made a video about Timnit's above tweet, calling it "disgusting" & He has 60k YT subscribers & https://t.co/u0lGa6vYV1 2020-12-23 16:18:08 @timnitGebru @beeonaposy @earnmyturns @JeffDean I didn't know that he worked at Google or that he'd made a video about this 2020-12-23 04:44:00 @BlakeleyHPayne @geomblog Thanks for searching Blakeley! Someone else just found it (although this has probably happened many times in the last few years): https://t.co/O8CoM9RIRg 2020-12-23 04:42:26 @jeggers Thanks for looking Jana! Someone else just found it: https://t.co/O8CoM9RIRg 2020-12-23 04:41:00 Thanks everyone! @beeonaposy found the thread. It was from April 2020, and I truly thought it was 2 or 3 years ago (yes, the last 8 months have been 3 years long) https://t.co/scUU6DcaY6 2020-12-23 04:37:31 @beeonaposy Yes, thank you Caitlin!! Although I can't believe it's from this April and I thought it was 2-3 years ago 2020-12-23 04:30:48 The discussion at the time highlighted how much dataset creation/collection is under-valued in ML research & 2020-12-23 04:16:49 @BlakeleyHPayne @geomblog Yes, I would guess maybe 2 years ago (give or take a year)? It was definitely a while ago, not in 2020. 2001-01-01 01:01:01