A summary of the draft paper co-authored by Timnit Gebru, which outlined the main risks of large language AI models and provided suggestions for future research

MIT Technology Review 2020-12-07 Karen Hao

Context & Ripple Effects

This summary lands at the flashpoint of Timnit Gebru's break with Google: her long-running critique of large language models had already put her at odds with the company, and the draft paper she co-authored turned that dispute into a formal research agenda — naming environmental costs, biased training data, and the inability of LLMs to understand language as core problems rather than edge cases.

What makes the document consequential is what grew out of it. Gebru left to found the Distributed Artificial Intelligence Research institute focused on responsible AI, and within months hundreds of researchers mobilized behind an open-source large language model built independently of any company — an explicit answer to the paper's complaint that corporate labs control both the models and the narrative about their risks.

First-order effects

Google and other corporate AI labs face direct pressure to address the paper's specific claims — energy use, dataset bias, and the limits of benchmark-driven evaluation — rather than framing LLM progress as neutral engineering.
Emily M. Bender and the paper's co-authors gain a shared vocabulary ('stochastic parrots'-style critique of scale-first research) that researchers and journalists immediately adopt in coverage of AI ethics versus AI safety.

Second-order effects

Independent alternatives accelerate: the 500-plus-researcher open-model project exists precisely so risk research doesn't depend on access granted by OpenAI or Google, shifting leverage toward academia and nonprofit consortia.
The paper's warnings get empirical follow-through — by 2022, MIT Technology Review is documenting [[a:982493|privacy concerns raised by GPT-3, LaMDA, Meta's OPT-175B and peers trained on personal data]], turning the draft's abstract objections into documented harms.

Third-order effects

If the pattern holds, LLM accountability migrates structurally outside the labs that build the models — via independent institutes like Gebru's DAIR, open-source model efforts, and external audits — forcing corporate labs to either open up or lose credibility as self-regulators.
The episode also hardens a personnel pipeline problem the relationships make explicit: prominent women and women of color in AI research leaving or being pushed out of Big Tech and building parallel institutions, which redistributes moral authority in the field away from incumbent labs.

The trend: Large-language-model risk assessment is shifting from internal corporate review toward independent institutions and open-source research infrastructure, with the Gebru controversy as an early catalyst.

Discussion

@math_rachel Rachel Thomas on x
I have long admired @timnitGebru for her brilliance, moral courage, clear voice in speaking up for what is right, & influential scholarship. It is truly terrible that Google would do this. In this thread, I want to share some of Timnit's work I love https://twitter.com/...
@william_fitz William Fitzgerald on x
This is such a lie. It was part of my job on the Google PR team to review these papers. Typically we got so many we didn't review them in time or a researcher would just publish & we wouldn't know until afterwards. We NEVER punished people for not doing proper process. https://tw…
@clairewaves Claire Stapleton on x
wrote something about the @timnitGebru ordeal having lived through something similar, and why the obvious answer to the question “why is google doing this??” just “because they are dumb” https://techsupport.substack.com/ ...
@_karenhao Karen Hao on x
If you're wondering what Google was so terrified of @timnitGebru publishing, I got a copy of the paper. Here's my best summary of what it says. Thank you to @emilymbender who coauthored the work for sending it my way. https://www.technologyreview.com/ ...
@jeffdean Jeff Dean on x
I understand the concern over Timnit's resignation from Google. She's done a great deal to move the field forward with her research. I wanted to share the email I sent to Google Research and some thoughts on our research process. https://docs.google.com/...
@lorakolodny Lora Kolodny on x
The controversial paper looks at: 1) the environmental footprint of advancing AI 2) how a certain approach to training data can perpetuate harm against any but the wealthiest and largest communities... (To me, this sounds like really worthwhile research.) https://www.technologyre…
@glowascii @glowascii on x
I support @timnitGebru, the #EthicalAI researcher ousted from Google. A paper she co-authored, which touched on her job, was subjected to extra, abnormal review, & after edits, she kept her integrity rather than retract. Sources: https://www.technologyreview.com/ ... https://www.…
@mer__edith Meredith Whittaker on x
RT @william_fitz: Google PR watches @Techmeme all day, so does senior mgmt. If a product launch gets top of the site for a few hours it's considered a real success. The firing of @timnitGebru has now been the top tech story for what is it now 4 days? https://twitter.com/...
@j_bothello Joel Bothello on x
“Gebru is one of the most well-known and respected Black female scientists working in AI.” Let me fix that typo for you @guardian: “Gebru is one of the most well-known and respected scientists working in AI” https://www.theguardian.com/ ...
@jeromobot @jeromobot on x
Really interesting insights about language and translation from paper that @timnitGebru @emilymbender et al wrote and that Google didn't like https://www.technologyreview.com/ ... https://twitter.com/...
@emmadaly @emmadaly on x
We read the paper that forced Timnit Gebru out of Google. Here's what it says https://www.technologyreview.com/ ... Great piece from @_KarenHao of @techreview https://twitter.com/...
@so_radhikal Radhika Radhakrishnan on x
Google keeps firing its best people, and that says a lot about the kind of organisation it is. In solidarity with @timnitGebru. I have great respect and admiration for her work. https://googlewalkout.medium.com/ ...
@whoseknowledge @whoseknowledge on x
We are huge fans of Dr. Timnit Gebru's ethical AI work. Last week, she was fired from Google for protesting the unconscionable censorship of research on environmental and ethical dangers of large scale AI language models. #SupportTimnit #BelieveBlackWomen https://googlewalkout.me…
@shiraovide Shira Ovide on x
I look forward to reading this research paper. A fascinating summary of it: https://www.technologyreview.com/ ...
@mcbyrne @mcbyrne on x
I hope advocates from the progressive space join me in solidarity for @timnitGebru. https://twitter.com/...
@tunisharpe Tuni Sharpe on x
Timnit Gebru , acting within her function at Google, used her expertise + beautiful mind to identify + challenge biases occurring in artificial intelligence. Now she doesn't have a job. Another Black truth-teller silenced via retaliation at work. Another reason we stay quiet. htt…
@_dmh @_dmh on x
Industry research, even when fairly open-ended, is not taking place in an “academic research unit”. It doesn't mean it can't be good research (good stuff came from AT&T, IBM, and Xerox PARC, etc), but I don't think we want these labs to absorb all the research talent either. http…
@teamcoworker @teamcoworker on x
“The #ISupportTimnit campaign, started by Google workers, accuses Google of ‘unprecedented research censorship’ and calls for transparency over why the paper was rejected and a commitment from the company to ‘research integrity and academic freedom’” https://www.nbcnews.com/...
@leesplez Jeff Lees on x
If I were a journal editor considering a manuscript from Google, I'd immediately halt the peer review process and demand from Google their formal policies regarding academic freedom after this episode. https://twitter.com/...
@rachelcoldicutt Rachel Coldicutt on x
So, the “didn't meet the bar for publication” argument either means this paper's authors were being held to a much, much, *much* higher standard than the rest of the research community, or that “the bar” is actually “the right answer according to Google”. https://twitter.com/...
@judywawira Judy Gichoya on x
Can tell you D and AI efforts are mainly about placing a bandaid on a big bleeding wound instead of cleaning and suturing the wound. there must be BRAVE Not SAFE places where people work and learn to move the needle further #IStandWithTimnit https://twitter.com/...
@alicialutes Alicia Lutes on x
I have a plot point about this sort of thing being brought to a boss' attention in a pilot of mine where a character literally gets fired for these same reasons!! This is so fucked!!! https://twitter.com/...
@emtiyazkhan Emtiyaz Khan on x
This wonderful thread goes over the two very problematic statements in Jeff Deans email. It is clear that the statements are vague in order to justify and hide the real facts. https://twitter.com/...
@carmeartigas Carme Artigas on x
According to @_KarenHao, @timnitGebru highlighted ethical AI risks on NLP systems as well as high CO2 emissions. Spain's AI Strategy #ENIA addresses these challenges; -National Green Algorithms Plan, NLP plan in spanish, Algo's Quality Seal and the Chapter of Digital Rights. http…
@lourdesgnavarro Lulu on x
Every word of this is worth a read . https://www.technologyreview.com/ ...
@wordglass @wordglass on x
You fired a good researcher and more importantly you really screwed up. https://twitter.com/...
@william_fitz William Fitzgerald on x
“These top Google execs may have power and with that some degree of control over livelihoods and billions of dollars and who knows what else. But that's not the same thing as respect and as this Timnit story makes excruciatingly clear: they don't deserve it.” https://twitter.com/…
@catacalypto Cat Manning on x
Good piece here on Google firing Dr. Timnit Gebru and what, among other outrages, a tremendous act of unnecssary self-destruction it is on their part. https://techsupport.substack.com/ ...
@omidyarnetwork @omidyarnetwork on x
“Your life starts getting worse when you start advocating for underrepresented people. You start making the other leaders upset” We stand by @timnitGebru, a leader in ethical AI who refuses to be silenced by @google. The public deserves greater diversity & transparency in tech. h…
@toyosirise Toyosi Akerele-Ogunsiji on x
“The team she helped build at Google is one of the most diverse in #AI, and includes many leading experts in their own right.” Interesting read. https://www.technologyreview.com/ ...
@mmitchell_ai MMitchell on x
“didn't meet our bar for publication ... It ignored too much relevant research”. He says of our paper, with over 128 cross-discipline citations. And so fires one of the authors. 2020's monster version of Reviewer #2. https://twitter.com/...
@willingcarol Carol Willing on x
An excellent thread about AI, research, and ethics. Resources created by/with @timnitGebru. Thanks @math_rachel for curating the impressive list. https://twitter.com/...
@helinnaayalew Helinna Ayalew on x
A thread on just some of @timnitGebru's contribution to AI and ethics research. A voice always willing to ask the hard questions & center marginalized ppl. If Google has no space for scholars like her, we should ALL be worried about how this tech will be used in the future. https…
@daniellecitron Danielle Citron on x
Thank you so much @math_rachel for this thread about the extraordinary influence and contributions of @timnitGebru — I'm hoping she goes into the Biden Admin but wherever she lands, we all benefit. Google we are watching—shameful and stupid you are. https://twitter.com/...
@evantachovsky Evan Tachovsky on x
Read this thread 👇 @timnitGebru is an agent of massive positive change. If the goal is to build safe and equitable algorithmic systems, we'll need more of her work, not less. https://twitter.com/...
@4sweb @4sweb on x
A thread, detailing some of Timnit Gebru's influential work, including GenderShade ( looking at error rates of facial recognition for women by shade of skin color), her ideas on bias, @black_in_ai , and more. https://twitter.com/...
@allisonkoe Allison Koenecke on x
.@timnitGebru directly inspired the fairness-related work in my thesis. There's only so much I can do as a relatively junior academic, but I'll no longer be interviewing for a postdoc with Google because #ISupportTimnit 1/ https://www.technologyreview.com/ ... 2/ https://googlewa…
@drbritwilliams Brittany M. Williams, Ph.D. on x
The best part of this story is them not realizing her leaving is why more of us know abt the paper. This prob would've been a big to do among AI, tech, & a few privacy academics / aficionados. But now it's mainstream BC a BW got fired for the thing she was hired for. Geniuses. ht…
@weskandar Wael Eskandar on x
This is such a well written article about how Google has pushed out a researcher for inconvenient truth about their AI approach. https://twitter.com/...
@lauren_marietta Lauren Chambers on x
I wonder how many young Black and brown folks in tech and academia are seeing this unfold and swearing off ever working at @Google. I know I'm on the list! If this is how they treat THEE Timnit Gebru, good God how would they treat us... https://twitter.com/...
@esouthershve Dr. Erroll G. Southers on x
Firing of high-profile Black female underscores Google's internal diversity battles. Timnit Gebru, technical co-lead of Google's Ethical Artificial Intelligence Team also said she felt “constantly dehumanized” at Google. https://www.nbcnews.com/...
@heathr Heather Gold on x
Important read and important paper. https://twitter.com/...
@mmadaio Michael Madaio on x
As per usual, @kharijohnson has some of the best reporting on AI issues, discussing here the paper that led to @timnitGebru's firing from Google, where the authors identify risks of large-scale language models https://twitter.com/...
@pkmital Parag K. Mital on x
. @Google how hard is it to not fire the person who is literally the voice of how not to be evil #IStandWithTimnit @timnitGebru If you haven't read or seen Timnit's work, check out this excellent conference as a start: https://sites.google.com/... https://twitter.com/...
@evanpeterwill Evan Will on x
Really interesting points about the difference between researchers in corporate and academic context. Of course meanwhile the good bits of academic context are being eroded... https://twitter.com/...
@blackamazon @blackamazon on x
I've come to those conclusions in media and journalism. I've seen in science, law, policy race .. and yet https://twitter.com/...
@lewis__bush Lewis Bush on x
This is very interesting read, not least because the observations in the paper seem entirely reasonable, and yet Google didn't think so (also omg check out the power costs and CO2 emissions for model training). https://twitter.com/...
@jennifermgmez1 Jennifer M. Gmez, Ph.D. on x
“Treatment of Exceptional Black Scientists” courtesy of Google - Attempt to silence (demand retraction of research) - Provide insufficient explanation (—> 'we don't value your expertise') - attack (w/o evidence, claim the work didn't meet your standards) https://www.npr.org/... h…
@lorinmackenzie Lori Nishiura Mackenzie on x
This is what an uneven playing field looks like. https://www.npr.org/...
@laurenceb @laurenceb on x
“It's not clear to researchers how they're going to continue doing this work in the industry,” said UC Berkeley computer science professor Moritz Hardt, who specializes in machine learning and has studied fairness in AI. “It's a chilling moment, I would say.” https://twitter.com/…
@shiringhaffary Shirin Ghaffary on x
My explainer on why Timnit Gebru's controversial departure from Google is setting off a firestorm in the tech and academic community. If you're at Google and want to talk, you can reach me via DM or shirin.ghaffary@protonmail.com https://www.vox.com/...
@iamcuriousblue Galway Curiousblue on x
Damn! Google actually fired someone for having controversial ideas about diversity in tech and making noise about it on their internal slack channel. What's the world coming to? https://www.vox.com/... https://twitter.com/...
@nrrrdcore Julie on x
Don't think I've read a better explanation re: NLP's inability to effectively detect nuance and/or keep up with constantly evolving language norms—the picture these models paint is wildly distorted, but profitable enough not to warrant scrutiny apparently https://twitter.com/...
@histoftech Mar Hicks on x
tbh this sounds like a brilliant paper and I hope we get the chance to see the whole thing soon https://twitter.com/...
@cattsmall Catt Small on x
“We are working at a scale where the people building the things can't actually get their arms around the data, and because the upsides are so obvious, it's particularly important to step back and ask ourselves, what are the possible downsides?” https://twitter.com/...
@amywebb Amy Webb on x
I urge you to read this important, nuanced, smart piece in @techreview by @_KarenHao about what went down last week with @timnitGebru. The researchers' concerns are exactly what I wrote about in my book on AI's problems The Big Nine. Short thread... https://www.technologyreview.c…
@bostonjoan Joan Donovan, PhD on x
Research should be an independent endeavor, but sadly industry has so many critical resources. We need independent and publicly funded research about the internet, AI, and computing now! https://twitter.com/...
@rossformaine Ross LaJeunesse on x
google used to tell employees to “act like owners” - i.e., speak up, ask tough questions, fix problems and challenge decisions - it's a key reason google's been so successful. now, speaking up gets you fired. #ISupportTimnit https://www.nytimes.com/...
@muna_mire @muna_mire on x
VERY troubling that Google's AI Ethics co-lead, Timnit Gebru, known for her work on AI and race (an interview guest when I was at @robinthede's show) has been forced out for raising the alarm on large language models! Read a summary of that research here: https://www.technologyre…
@rosejackson3 @rosejackson3 on x
We have to start talking about the subtle-& not so subtle-hidden ways the tech industry makes it hard, then wounding, then impossible for women & people of color to succeed. @timnitGebru is but one GLARING example. But anyone in tech knows someone or has experienced it themselves…
@browserdotsys Bowser on x
imagine getting fired over a mishmash of hot topics that have been circulating for the past few years https://twitter.com/...
@jackclarksf Jack Clark on x
An analogy of this situation - imagine if particle physics was solely domain of private sector and @timnitGebru had a critique about some societal effect of running atom smashers - how would we feel if the company suppressed that research? Bad, I think!
@jackclarksf Jack Clark on x
I find it notable that the paper was about these models that have been developed in private industry (GPT3, BERT, etc). But what the paper is pointing at is the need for a far greater amount of study applied to these systems. That requires massive engineering resources.
@miles_brundage Miles Brundage on x
Hard to imagine a way that Google could have called more attention to this paper.
@juliaxgulia Julia Kaganskiy on x
Perhaps the most surprising thing about the research paper Google suppressed and fired Gebru over is how well established its contents appear to be. AI's carbon footprint, tendency towards bias, and consolidation of power in the hands of major corps is not exactly news, is it? ht…
@omanreagan Michael on x
Corporations can't be in charge of and be gatekeeper to the research that examines the impact of what they do. This is proof of that at Google. Read it. https://www.technologyreview.com/ ...
@jtasioulas John Tasioulas on x
“Many of the top experts in AI ethics work at large tech companies because that is where the money is.” Interesting article. https://www.technologyreview.com/ ...
@azeem Azeem Azhar on x
the inconvenient truth seems to be that @timnitGebru & others highlighted real problems of prejudice and bias in the large scale language models Google is rolling out for its key search product. Cc @JeffDean / Cracking work from @_KarenHao cc @emilymbender https://twitter.com/...
@rachelcoldicutt Rachel Coldicutt on x
This is vital research for everyone deploying services based on large-scale language models - not just Google but governments, charities, businesses. Efficiencies aren't efficient if they deepen structural bias. Thanks @timnitGebru for doing, and standing by, this important work.…
@catherine_mayer Catherine Mayer on x
The best explainer yet on why Google forced out the amazing @timnitGebru - and how important her work is https://twitter.com/...
@_karenhao Karen Hao on x
@MichaelSmithLL @mer__edith @timnitGebru @emilymbender @JeffDean Hi Michael, here's a good account from @tsimonite at @WIRED. Timnit tried to work with Google on a solution but management refused. https://www.wired.com/... https://twitter.com/...
@ngkabra Navin Kabra on x
“Stochastic Parrots” is a great name for all machine learning (as opposed to artificial intelligence). Apparently, the paper that @TimnitGebru was fired over has a title that starts with “On the Dangers of Stochastic Parrots: ...” https://venturebeat.com/...
@dseetharaman Deepa Seetharaman on x
This paper sounds important & fascinating, especially the part about the near-impossibility of auditing these systems for bias because of their size. “A methodology that relies on datasets too large to document is therefore inherently risky.” https://twitter.com/...
@aprilaser April Glaser on x
One former employee who the NLRB ruled this week was unlawfully fired after labor organizing at Google, @laurenceb, had incisive comments in response to Dr. Gebru's firing https://www.nbcnews.com/... https://twitter.com/...
@eileenguo Eileen Guo on x
“We are working at a scale where the people building the things can't actually get their arms around the data,” says coauthor @emilymbender. https://twitter.com/...
@lisawshapiro Lisa Wood Shapiro on x
from @_KarenHao “Gebru's draft paper points out that the sheer resources required to build and sustain such large AI models means they tend to benefit wealthy organizations, while climate change hits marginalized communities hardest.” https://www.technologyreview.com/ ...
@superwuster Tim Wu on x
On Google firing Timnit Gebru — the whole idea of a vague “internal review” that Google researchers need pass crazy to me (if for anything but trade secrets/ IP) Risk of censoring research seems high. I'm trying imagine if Columbia U had that... https://www.technologyreview.com/ …
@techreview @techreview on x
We got a copy of the research paper that forced Timnit Gebru out of Google. In it, the company's star ethics researcher highlighted the risks of large language models, which are key to Google's business. https://www.technologyreview.com/ ...
@frankpasquale Frank Pasquale on x
“Though most AI researchers acknowledge that large language models don't actually understand language and are merely excellent at manipulating it, Big Tech can make money from models that manipulate language more accurately, so it keeps investing in them.” https://twitter.com/...
@celiahodent Celia Hodent on x
All seems to be going well in terms of #inclusion at Google, as well as #ethics regarding their #AI work... What kind of a surreal move is this? https://twitter.com/...
@jenstirrup @jenstirrup on x
I put my career on the line by handing back the #MVPBuzz Microsoft MVP Award after harassment and bullying. It was a scary time but I could not continue being diminished anymore by takers who stood silent about my treatment. I feel for @timnitGebru today. https://twitter.com/...
@daiwaka Daisuke Wakabayashi on x
36 hours after @timnitGebru went public with her departure from Google, here is the company's first (carefully-lawyered) public comment. You can see from the replies that it's not being received well. https://twitter.com/...
@lexialex @lexialex on x
This is a powerful white man gaslighting a Black woman. Not cool. She said she was fired. I believe her. https://twitter.com/...
@blackamazon @blackamazon on x
This email doesn't reference any of the research it believes is more efficacious versus Timnit's . It provides no counter to Timnit's timeline of events besides implying she was nit supposed to ask despite being head of AI The fact you sent this email doesn't answer concerns http…
@novallswift Novall Swift on x
Hi all. Apple has growing efforts in AI Ethics and we care about doing what's right. I've made personal connections with folks who are hiring in AI/ML and I would love to speak with anyone who worked with Dr. Gebru who might be looking for a change. My DMs are open.
@ainowinstitute @ainowinstitute on x
We stand with Dr. Timnit Gebru, and are grateful for her brilliant and tireless work. If you want to show your support, Dr. Gebru's colleagues have drafted this petition: https://googlewalkout.medium.com/ ... #IStandWithTimnit
@4sweb @4sweb on x
We stand with Dr Timnit Gebru, who was fired from her position at Google, after she co-authored a research paper that raised concerns about Google's AI/language models. This matter is a question of academic integrity, censorship, and Google's commitments to DEI. https://twitter.c…
@maddiestone Maddie Stone on x
Google is better with @timnitGebru. Googlers, please consider signing this letter in support of Timnit. It matters. https://googlewalkout.medium.com/ ...
@kharijohnson Khari Johnson on x
“In summary, we advocate for an approach to research that centers the people who stand to be affected by the resulting technology, with a broad view on the possible ways that technology can affect people,” the paper reads. https://venturebeat.com/...
@marilink Mariln Gonzalo on x
“The research paper surrounding Gebru's exit questions the wisdom of building large language models and examines who benefits from them, and who is impacted by negative consequences of their deployment” https://venturebeat.com/...
@catherineadenle Catherine Adenle on x
Seriously?! The day Google was found to have illegally spied on and terminated activists was the day they fired their top #AI #ethics expert who is a black woman. Poor judgement/PR. #diversitynotdivision #DiversityMatters #bias #WIT #WomenWhoCode #inclusion https://twitter.com/..…
@oliviasolon Olivia Solon on x
The research paper that Google said @timnitGebru had to retract looked ethical problems with large-scale machine learning models including their environmental impact and potential to encode biases. Google uses one such model to power its search engine. https://www.nbcnews.com/...
@grady_booch Grady Booch on x
I look forward to someday reading @timnitGebru's research paper. https://twitter.com/...
@biosbenk David Benkeser on x
This whole saga is hugely disappointing, but not unexpected. Tech has opened up a lot of great research opportunities for quantitative researchers. One thing those jobs lack that's critical for research in sensitive fields? Tenure. https://twitter.com/...
@caseynewton Casey Newton on x
Was talking to a manager yesterday about the Timnit Gebru situation and they said that if an employee threatens to resign in writing, employers can legally treat that as a resignation. Anyone else hear this / have experience here? Feel free to DM
@nitashatiku Nitasha Tiku on x
There's much to say & learn still about Google firing a pioneer like @timnitGebru, who gave its AI ethics credibility you just can't replicate. But it's worth focusing for a min on the fact that the email she sent was to an ERG for women and allies in AI https://www.washingtonpos…

Chronicles