Timnit Gebru on her sacking by Google, AI’s dangers and big tech’s biases

https://www.theguardian.com/lifeandstyle/2023/may/22/there-was-all-sorts-of-toxic-behaviour-timnit-gebru-on-her-sacking-by-google-ais-dangers-and-big-techs-biases

by biorach • 3 years ago

29 61 3 years ago

theguardian.com

curiousgal3• 3 years ago

The Ethiopian-born computer scientist lost her job after pointing out the inequalities built into AI.

I am glad they included that at the start of the article so I know to completely skip reading it!

ggm2• 3 years ago

You're doing yourself no favours. Even if you disagree with her you should inform yourself of her line of reasoning.

I think the guardian article is a reasonable write up of the issues.

curiousgal3• 3 years ago

I don't disagree with her. My gripe is wit the the article itself. If the reporter openly misrepresents (read lies) about obvious facts, I don't see how their take on more nuanced issues would be..nuanced and unbiased.

She quit/was fired for submitting a paper without following company procedure. Saying she lost her job after pointing out AI ethics issues paints a completely different picture and is not productive at all.

prepend3• 3 years ago

lost her job after pointing out the inequalities built into AI.

I think that this is a simplistic and incorrect summary of why she no longer works for Google.

It would be as accurate as if the article started with “lost her job after violating Google’s publishing clearance procedures and sending out inaccurate and accusatory emails to internal listservs”

surgical_fire4• 3 years ago

I read the article. While it does mention the keywords promised on the title, it's also very long winded as it recounts many of her personal life experiences.

Honestly, a very boring read. Brings nothing new to the discussion, and at times feels just shy on gushing about the subject matter.

As for her being fired from Google... Google being Google, thet're probably in the wrong. But it is not surprisingly in the slightest that this sort of activist employee ends up being sacked when they start to be troublesome in their activism. Hell, I wonder why Google even hired her in the first place. My guess is that they hoped that by receiving a paycheck, she would be neutered.

rg111• 3 years ago

Her life's biggest highlight is being sacked by Google whom she had given an ultimatum which to me was unfair.

aaron695• 3 years ago

Ethics employees are the lowest on the Bullshit jobs scale.

They bike shed all day. Just regurgitate old stories and what ifs.

They never do anything quantifiable, they don't have the technical skills.

Unlike middle Bullshit jobs that do nothing, they instead drag down employees around them.

Case in point here and as we can see Google is currently losing the AI battle.

mempko5• 3 years ago

Except Timnit Gebru did quantifiable research on AI bias and was punished for it.

RcouF1uZ4gsC• 3 years ago

Here is Jeff Dean's email about the situation when it happened

https://www.platformer.news/p/the-withering-email-that-got-a...

Hi everyone,

I’m sure many of you have seen that Timnit Gebru is no longer working at Google. This is a difficult moment, especially given the important research topics she was involved in, and how deeply we care about responsible AI research as an org and as a company.

Because there’s been a lot of speculation and misunderstanding on social media, I wanted to share more context about how this came to pass, and assure you we’re here to support you as you continue the research you’re all engaged in.

Timnit co-authored a paper with four fellow Googlers as well as some external collaborators that needed to go through our review process (as is the case with all externally submitted papers). We’ve approved dozens of papers that Timnit and/or the other Googlers have authored and then published, but as you know, papers often require changes during the internal review process (or are even deemed unsuitable for submission). Unfortunately, this particular paper was only shared with a day’s notice before its deadline — we require two weeks for this sort of review — and then instead of awaiting reviewer feedback, it was approved for submission and submitted.

A cross functional team then reviewed the paper as part of our regular process and the authors were informed that it didn’t meet our bar for publication and were given feedback about why. It ignored too much relevant research — for example, it talked about the environmental impact of large models, but disregarded subsequent research showing much greater efficiencies. Similarly, it raised concerns about bias in language models, but didn’t take into account recent research to mitigate these issues. We acknowledge that the authors were extremely disappointed with the decision that Megan and I ultimately made, especially as they’d already submitted the paper.

Timnit responded with an email requiring that a number of conditions be met in order for her to continue working at Google, including revealing the identities of every person who Megan and I had spoken to and consulted as part of the review of the paper and the exact feedback. Timnit wrote that if we didn’t meet these demands, she would leave Google and work on an end date. We accept and respect her decision to resign from Google.

Given Timnit's role as a respected researcher and a manager in our Ethical AI team, I feel badly that Timnit has gotten to a place where she feels this way about the work we’re doing. I also feel badly that hundreds of you received an email just this week from Timnit telling you to stop work on critical DEI programs. Please don’t. I understand the frustration about the pace of progress, but we have important work ahead and we need to keep at it.

I know we all genuinely share Timnit’s passion to make AI more equitable and inclusive. No doubt, wherever she goes after Google, she’ll do great work and I look forward to reading her papers and seeing what she accomplishes.

Thank you for reading and for all the important work you continue to do.

-Jeff

user_named• 3 years ago

Classy response to a not so classy person.

DiggyJohnson• 3 years ago

This is honestly a very respectable and transparent communication.

lakomen• 3 years ago

"AI threatens to deepen the dominance of a way of thinking that is white, male,"

I stopped reading there.

I'm white I'm male I have 0 advantages because of that. Always the same lies everywhere. I'm tired of it.

giraffe_lady3• 3 years ago

You have to be able to see the problem with this reasoning right?

https://twitter.com/nathanwpyle/status/1031008855210123264

surgical_fire4• 3 years ago

I'm not white.

I don't believe that my "way of thinking" is meaningfully different from the way of thinking of a white person, language and cultural differences accounted for.

Actually, I firmly believe that implying that is profoundly racist. It's small wonder that I see most dangerously racist bullshit being parroted in progressive circles.

askin4it2• 3 years ago

Do you see the problem with your cartoon tutorial. Correct me if I'm wrong but your lesson is a good example of begging the question . I think of the two fallacies, the whipped white guys is not nearly as bad as yours. His is merely anecdotal. Yours rhetorically casts that sad sack as a predatory animal.

Please stop helping.

endisneigh• 3 years ago

I'm white I'm male I have 0 advantages because of that. Always the same lies everywhere. I'm tired of it.

Even if it were true how would you know?

tanseydavid• 3 years ago

> Even if it were true how would you know?

I think a better question would be "How would YOU know?"

tr_user2• 3 years ago

They mean to say patriarchy and white supremacy. The fight is against the system. I don't know why so many people don't use better terminology that doesn't exclude allies.

barbacoa• 3 years ago

It is a little ironic to think that people being patronaged by giant monopolistic mega corporations are fighting "against the system".

SilverBirch6• 3 years ago

Well if you stop reading any time anyone suggests you might have advantages you're not aware of then it's not surprising you don't think you have any advantages.

onetimeusename• 3 years ago

is it possible to disprove that "AI deepens the way of thinking that is white, male"? I don't think it is for obvious reasons. It's perfectly reasonable to stop reading when you encounter an argument like that.

thefz• 3 years ago

I remember at the time Gebru basically went into leave until all her complaints were addressed, which can be considered as soft quitting for a company like Google.

FooBarBizBazz2• 3 years ago

There's a funny thing about this claim. These models are now trained by a kind of reinforcement learning with feedback from humans. Who are those humans and where are they?

https://slate.com/technology/2023/05/openai-chatgpt-training...

SilverBirch6• 3 years ago

I find it fascinating that in the middle of a hype cycle here, where really central people to AI are saying we should be worried (Altman, Hinton etc.), there's still just outright animosity towards people who are actually interested in studying these problems. It's also very far from a theoretical problem, did you know that crash test dummies are men? Literally, until extremely recently crash test dummies were all designed to be 50th percentile men[1]. Surely they test women dummies too? Oh yeah! They put a small woman dummy in the passenger seat. But still, they've presumably measured women and designed something appropriate? Nope! The small female dummy is defined as a scaled version of the male dummy. It's genuinely astonishing how short sighted engineers can be in designing technology and it's wild how angry they get when anyone tries to actually fix it.

[1] https://www.sciencedirect.com/science/article/pii/S000145751...

mempko5• 3 years ago

Exactly. Same problem with facial recognition which simply didn't work on anyone with dark skin. The bigger the models the less obvious the bias becomes but is still there. Gebru did good work quantifying that but google shut it down.

And it's sad that many of the comments here are reactionary and hostile towards Gebru. Why is the community here so bad? I thought it's supposed to be full of smart people with computers?

908B64B1974• 3 years ago

hostile towards Gebru

Her paper aren't really high quality. Her carbon footprint calculations had all of Google's compute running from jet fuel. [0]

In her "Gender Shades" paper she completely ignored the existence of the Asian race. That's a pretty big limitation for a paper named in such general terms. In another place she labels Asians as "white adjacent".

She can't argue about technical subjects (here's an example [1]) because she isn't technical.

[0] https://www.technologyreview.com/2020/12/04/1013294/google-a...

[1] https://syncedreview.com/2020/06/30/yann-lecun-quits-twitter...

revelio6• 3 years ago

Same problem with facial recognition which simply didn't work on anyone with dark skin.

That's an urban legend btw. The issue was actually that the person for whom it failed had the camera pointed straight up with a bright light behind him, so the image was totally washed out and lacked saturation. The software failed in the same way if you had a dark background and then shone a bright light into a white person's face. The algorithm needs contrast to work.

Also, people on this forum really should know better than insinuate that a program working imperfectly is caused by racism. As if programmers sit around finding ways to troll black people by deliberately making face recognition not work. Even the idea of it is absurd. When people make such insinuations, it reinforces the widely held belief that accusations of racism are invariably bullshit, which can lead to a boy-crying-wolf problem.

dragonwriter• 3 years ago

I find it fascinating that in the middle of a hype cycle here, where really central people to AI are saying we should be worried (Altman, Hinton etc.), there's still just outright animosity towards people who are actually interested in studying these problems.

Note that the abstract speculative future concerns Altman and Hinton are concerned with are very different from the immediate and present issues Gebru, her former Google colleague Margaret Mitchell (now at Huggingface), and others in the AI ethics space are raising.

SilverBirch6• 3 years ago

Absolutely, I think actually her concerns are much more direct, measurable and actionable. It didn't take a genius to point out "maybe we should crash test something other than a male crash test dummy", but it was much more impactful and actionable than "let's figure out how to test self-driving cars" (to which the answer is still "let's just drive them and hope")

rayiner• 3 years ago

The thing that was very confusing to me as an immigrant was that liberal type of racism,” she says. “People who sound like they really care about you, but they’d be like: ‘Don’t you think it’s going to be hard for you?’

This is why I object to teaching kids about racism in schools. Maybe the teacher was racist. Or maybe she got good grades—which isn’t hard to do in American high schools—but her teachers didn’t think she was one of the exceptional kids who could go on to excel in the field as a career. If you aren’t socialized to think of yourself as the target of racism, the natural response is to probe further. “But my grades are good!” The explanation that comes next might lead to learning and growth, or at least renewed motivation (“I’ll show him!”) But if a kid has been socialized to perceive such interactions in terms of racism, there are, they can’t learn from such conversations. After all it’s pointless arguing with a racist, nor can you learn anything valuable about yourself from a racist.

I moved to Virginia from Bangladesh back in the 1980s, to a school and neighborhood that was nearly 100% white. At the time, northern Virginia where we were was solidly Republican. But my parents never taught me about “racism” and it was a tremendous blessing. I never wondered whether the ordinary negative interactions of life were racially motivated. Was that kid being a bully because I was brown, or because I was a book worm? What good would it have done me to recognize it anyway?

I’m extremely worried about how the discourse is going to affect my kids. It’s impossible to protect your kids these days from folks like Gebru.

prepend3• 3 years ago

‘Don’t you think it’s going to be hard for you‘

I found this curious when reading the article and expected more context as to why this was racist. Why did Gebru consider this was due to her race or immigrant status? I got this comment through my whole life too even though I don’t share Gebru’s background or demographic factors. Were my teachers racist? Or were they concerned? Or was the subject just really hard? Or were they classist against my poverty?

It seems more info is needed and attributing this to racism (or sexism or xenophobia) seems like an unsupported conclusion. Although it seems Gebru felt it was racist and significant enough to mention in the background for this article.

I think this is what makes measuring micro aggressions so difficult. It’s rare to confirm the intent of the person who did something that is perceived as micro aggressions but the recipient feels and infers an intent.

I don’t know how to fix this but I want to.

I remember once remarking to a coworker that they had a nice car. It was a new 7-series and looked really nice. I thought nothing of it as it was just small talk and I frequently make such small comments to note something positive.

A few months later we were having a large zoom call after the George Floyd murder where people were recounting persistent and systematic racism. My coworker shared an example of what he called everyday racism when someone said he had a nice car and he assumed it was because he was black and couldn’t afford such a nice car. He seemed really hurt and spoke for 5 minutes about it and how he worked really hard to afford the car and his spouse had a high income as well. He didn’t name me as the person who made the comment but it sounded like my small comment.

I was so confused as to how all that could be inferred from “nice car.” I don’t know him well enough to ask about it and clear it up. I felt bad he was hurt, even though I just intended it as a small comment recognizing a nice car from a coworker and would have made the comment regardless of his race. So I don’t think I did anything wrong but still feel bad. And there’s no way I could predict such a response.

I don’t compliment coworkers cars any longer.

netsharc• 3 years ago

I don’t compliment coworkers cars any longer.

It's unfortunate, but for you to feel guilty and to stop this seems unreasonable. There are myriad of traps I can imagine you'd be afraid of falling into. Like, "Oh I want lunch and I fancy some fried chicken. Oh wait if I invite [black colleague] will they be offended at the suggestion to go for that food?"

Communication goes through filters, as this cheap graphic shows, and I don't think it's something you can fix on your side:

https://external-content.duckduckgo.com/iu/?u=https%3A%2F%2F...

david38• 3 years ago

Wow. Totally wrong lesson learned. Go refresh yourself to some hardcore problems and then replay your situation again.

A more standard response may be: “oh that was me. Your car is better than mine. Sometimes you know, a compliment is just a compliment.”

I’m currently reading Hunting Eichmann. Really nasty stuff that went down, as many know. A very strong reminder that so many of these first-world problems, micro-aggressions, sympathy-seeking woe-is-me public sob stories are just crap. Seriously?? He felt bad from a complement about his car??

lisasays2• 3 years ago

So I don’t think I did anything wrong but still feel bad. And there’s no way I could predict such a response.

I feel bad for both of you. This anecdote seems to capture everything wrong with the current climate of discussion on these topics in the US right now.

908B64B1974• 3 years ago

is perceived as micro aggressions but the recipient feels and infers an intent.
I don't compliment coworkers cars any longer.

I'm starting to think these "micro aggressions" are just a way to keep a culture war going, because some groups need a cause to fight for (for political reasons).

It reminds me of the whole master vs main debate.

A (black) engineer colleague of mine told me about his team's effort to change master to main. The whole initiative was started by a rainbow colored hair (white) PM and since it was what they believed to be a highly visible and easy fix, grew to a team of 5. All non-technical PMs of course. They ended up producing a "manifesto of inclusive software" where they listed every word they considered offensive and what it should be replaced with and made a very public announcement regarding the change.

The only response to their email was my (black) colleague asking if the branch renaming could be postponed to after a release because he didn't know what it could break in the build and release automation in case "master" is hard-coded somewhere.

This apparently started a lengthy thread between him and the 5 PMs where they explained to him that the reason he wasn't supportive of the change was because of the "systemic and cultural racism" he apparently internalized.

FooBarBizBazz2• 3 years ago

Large models are indeed dangerous, because they threaten to concentrate power in the hands of a small number of tech companies with the resources to do the training, and because, once those models exist, they devalue human labor.

Some technologies are leveling. It may be debatable, but I've heard it argued that the musket was one of these. The musket was cheap. It rendered expensive mounted plate-armored aristocratic warriors obsolete. Down came the whole feudal hierarchy necessary to keep those horses fed.

These AI models are the opposite. They do not spread power around; they centralize it. They increase dependence.

So in many ways she's right.

What I would like to ask, though, is -- instead of deepening the intersectional cuts that were used to dismember Occupy, what can a technologist do to both reduce centralization and increase solidarity? Is it possible to have a broad-based populism that isn't particularist? What technologies could be in its arsenal?

Possibly AI (which is really compression) can be part of this if it enables offline, disconnected "compressed Internet"s to exist off-the-grid, like solar panels?

How do we reduce dependence on the platforms? The different ethnic factions wouldn't fight each other if they were more self-sufficient. If they weren't made to compete with each other for access to a small number of economic chokepoints. How do we destroy the chokepoints? "Fuck your Suez, your Panama -- we have airplanes."

nh23423fefe• 3 years ago

After her departure, Gebru founded Dair, the Distributed AI Research Institute, to which she now devotes her working time. “We have people in the US and the EU, and in Africa,” she says. “We have social scientists, computer scientists, engineers, refugee advocates, labour organisers, activists … it’s a mix of people.”

Virtue signal by donating to my slush fund.

revelio6• 3 years ago

Amazing that these people don't even pretend. Refugee advocates and labour organisers have what to do with AI research? Nothing? Like La Sombrita and how the attempt to design a bus stop somehow became connected to female empowerment. The left asks for money to do X and then happily and very publicly spend it their usual ideological goals, even though they have nothing to do with X.

Wonder when this crosses the line and becomes some sort of fraud. Did they tell people they'd spend grants on AI research? A lot of their stuff seems very close to political advocacy and lobbying. They flat out state they hired activists and advocates. A 501c3 is supposed to have tight controls on political operations though.

Timnit Gebru on her sacking by Google, AI’s dangers and big tech’s biases

Related Stories