Google CEO pledges to investigate exit of top AI ethicist
axios.com
axios.com
1 - Google fired Gebru for trying to push a paper that presented google in a negative light. Doing so by creating special magic policies for her and then telling her she was fired. The concern in this case appears to be the fact that Google is suppressing criticism and ignoring it's own Ethics department and this is an example of morality being destroyed for profit.
2.- Gebru acted in an incredibly unprofessional manner following instruction to retract a paper, including posting charged complaints and accusations on an internal company mailing list as well as demanding to have revealed to her the identities of people that reviewed her paper and found it wanting. Those in this camp seem to generally accept that her firing was a result of unprofessional behavior from an employee that was disruptive to the organization. They then see that Google accepted her offer of having her demands met or working on an exit date as a reasonable business decision [See note], and see the subsequent drama as an attempt to draw attention to herself. Ultimately they see it as removing a toxic employee.
I don't really know what my point is but it seems in these comments we generally have 2 totally different discussions going on with one side lambasting the other for being ignorant, naive, or clueless.
Note: Also point of clarification that I don't see often asserted, it is SOP to remove a disgruntled employee's access to everything as quickly as possible after (or often before) termination in order to prevent them from using their access in a manner detrimental to the company.
Google had been wanting to get rid of Gebru for a while, for both good reasons (based on her emails and tweets she does indeed seem like an incredibly toxic personality and I sure wouldn't want to work anywhere near her) and bad (retaliation against the lawsuit she filed last year, which is both unlawful and incredibly unethical). They picked this paper as what must have seemed like a clean reason when she made the threat to resign, but taking her up on that threat has blown up in their face because everyone can see their supposed rationale doesn't hold water.
Taken as a whole, neither side looks particularly sympathetic, which makes it easy for partisans to latch onto the bad behavior of whomever they dislike.
I also don’t believe that Google would have done this for no reason (given her behaviour during the LeCunn incident) so I’m sympathetic towards Google’s conduct. If Gebru had simply come out and said “Google tried to censor my paper and so I was forced to resign”, I would be thinking very differently.
It still doesn't translate to a firing.
If you resigned, generally you can't, barring some specific circumstances.
Either side can terminate the employment earlier.
I'd assume best intent that she would not go with something like "my last day is a year from now".
Why would if she leave immediately or 14 days later? If she didn't get to wrap up her work, then that's Google's loss.
To be honest, trying to spin this as Google firing her is very very strange. It's a non-issue, and she did throw an ultimatum.
Toxic to management, embarrassing to her line managers or toxic to the colleagues she co-produces her work with?
Was she toxic to the company's profits, toxic to the company's profits, or both?
You can have the loyalty of a couple of close coworkers and still be toxic for the company as a whole, if you don't have professional interactions with other teams, or with your own managers.
Stereotyping is doing a lot of work here, driving people's assumptions about Gebru.
Where I work, if someone is ousted, people might talk about the person’s toxicity or whatever got them fired one on one or in small groups, in a pretty off the record kind of way.
I’d bet the people who agree with the petition think it’s representative and the people who disagree think it’s not representative (“people are generally like me” is a reasonable prior).
1. If someone is to sign a petition for Gebru, what could they lose?
2. If someone were to speak out, under their name, against Gebru, what could they lose?
3. If someone is neutral on the issue and a pro Gebru petition gets passed around asking for support, what is the path of least resistance?
See also https://www.reddit.com/r/MachineLearning/comments/k77sxz/d_t...
I agree I wouldn't want her reporting to me, but is it really that hard to fire someone in California? From the outside Google management seems to endorse solidly left-leaning politics, so while anything is possible it seems unlikely that they objected to her views or activism. If they think she is toxic, it follows that they likely had actual reasons - and if there are actual reasons someone is a bad employee then they can be fired in my experience.
That lends some credence to the retaliation theory. If they can't expel her for good reasons, grab the first half-baked opportunity.
I had an experience once where a toxic, delusional, pathological liar of an employee was hired and started making up wild accusations about their coworkers on day 1. They then proceeded to do no work and had their role at the company progressively minimized so that they could have satisfactory performance for how little they were doing (and couldn't do much damage). There was no way the company was going to fire them and thankfully after a couple of years they decided to quit, despite making their entire team miserable that whole time.
I think it has to do with lawsuits being highly visible whereas underperforming employees are sort of an invisible cost. So, the risk assessment focuses on the rare but highly visible occurrence rather than the ongoing cost of continuing to employ someone who isn’t productive.
Just guessing. I don’t really understand this phenomenon.
And yet, my perception is that companies often resist terminating problem or unproductive employees due to the perceived risk of being sued.
Discriminating against a member of a protected class is against the law, so perhaps they are worried about being accused of that?
That seems like common sense.
To be untouchable be openly degenerate & filthy rich.
Are you thinking of any specific case where the Times doxed someone for practicing intellectual freedom?
Her behavior fits the definition of 'toxic employee', it was unprofessional by any standard.
Here is Jeff Dean's letter sent to the Google Research team with the details of what actually happened: :https://docs.google.com/document/d/1f2kYWDXwhzYnq8ebVtuk9CqQ...
You choose to trust the words of the man that did the firing vs. the corroborated stories of dozens of Google Brain researchers that contradict his corporate messaging. Don't take corporate's word as "what happened:" that's about as naive as taking a police's side as the "truth."
It's a "boy who cried wolf" for me. I've seen these activists promoting lies to us the entire summer of protests for completely made-up causes. I've seen them distorting facts and pushing misinformation through media. I've also seen them bullying people via social media 'witch hunts', which is happening now with Jeff Dean, Nando de Freitas and whoever stands in their way.
The whole activism movement is toxic and whoever brings it to the workplace.
Why do you find people that are putting in the work to bring fairer, more just, and more even-handed application of human rights to all corners of society "toxic"?
What are you afraid of, or believe is "toxic", by removing the gap in privilege* that exists between white men and everyone else?
What I see is a bunch of bullies forcing their opinions on everyone else, spreading lies and propaganda on media, attacking and 'cancelling' anyone who stands in their way.
Whatever seed of 'social justice' motivation was there, the activist movement completely discredited themselves, and the ideas they so aggressively promote. And the more I interact with them the more it seems their primary motivation is promoting themselves.
In context, it only makes sense. Google is not going to admit in a memo that they have a structural problem with either race or ethics, but it still is ever so slightly disappointing to see this treated more or less as an internal PR issue and not a moment to look in the mirror.
The real mismatch here is what people think the core issue is vs what Google thinks the core issue is. From the memo, Google seems to think that some processes or policies would have prevented this. What it seems to me is that the general public sees a misalignment / direct conflict of values between Gebru's work and points and Google's profit / business. I think Google knows this internally but can't say the quiet part out loud.
I don't know if Gebru actually qualifies aside the matters of ethical AI. In my experience those people who work in engineering who sidewalks into some non-engineering matters often lack skills to compete where most do, and decided to leverage their incompetency with some other topics. And Gebru is doing exactly that in Google.
As I understand it Gerbu was hired as an academic to do research into AI ethics (her field of study). She was fired, in part, due to a disagreement around the contents of a paper she was publishing.
Where is the sidewalking into engineering matters?
If you've been in academia for awhile you've seen this happen time and time again. Student does excellently in undergrad, masters, phd coursework but when it comes to making real technical contributions, they hit a wall. It was a clever move by her to use her pedigree to pivot into "AI ethics" where the standards for publishing are much lower (as we can see from this draft that was circulated).
In general? Yes. When it comes to the snakes FAANG have created biting the hand that feeds on the industry stage? No. I absolutely love to watch their internal dramas spill out for all to see or, if you're smart, take warning from.
>I don't know if Gebru actually qualifies aside the matters of ethical AI. In my experience those people who work in engineering who sidewalks into some non-engineering matters often lack skills to compete where most do, and decided to leverage their incompetency with some other topics. And Gebru is doing exactly that in Google.
Not in a position to comment on this person's credentials or skills but I don't personally begrudge the type of grifter you're talking about. Most of us are merely trying to optimize for income and they're no different, and their success is a result of poor leadership in companies and/or society.
Looking at this Twitter thread between her and Jeff Dean 6 months ago: https://twitter.com/JeffDean/status/1278571537776271360 ...
She was being toxic toward Jeff Dean, who is her manager's manager.
I looked at that tweet, and yikes. I'll never do that to my colleagues.
However, I think it's much more likely her toxic behaviour is a direct result of her being put into an extremely difficult position by her employer. Her job is to publish research on the ethics of the AI work that Google is doing. But by the sounds of it, Google want to block her from actually publishing anything critical. It would be one thing to say that her work was unsound, but that doesn't really seem to have happened. Instead there seems to be the classic corporate pocket veto. They don't want her publishing something critical, but they also want the PR of having respected AI ethics researchers. So they don't say she can't publish critical work, they just won't let her publish critical work. And when she asks why she can't publish critical work... well she can publish critical work Google has always been very clear that she's free to publish critical work. So she can publish the work? No. No you can't publish the work. That would make me pretty crazy too. I don't she's handled it well, but at the end of the day she's a person and she's facing her job basically being taken from her (and arguably finding out she's been a useful idiot for quite a while).
It's pretty clear here that Google wants to have its cake and eat it too. It wants to tell the world that it's got respected scientists doing research into the ethics of their AI work, but they don't actually want researcher to do research about the ethics of their AI work because a lot of what they're doing is unethical.
Have no doubt: this CYA text was massaged and went through several HR and in-house GC revisions, and maybe even a few round-trips to outside labor litigation counsel.
In other words, this statement is pure CYA and litigation-defensive. They're terrified of a racial discrimination lawsuit.
I hope that Timnit can find her way to the courthouse. I'd love to cross-examine Pichai with this statement in hand. He would deny that these were his words. He would say that he would have expressed it differently. He would say that he wasn't involved with the firing decision: every time the question is asked, the noun used is "termination" and not departure.
The only question remaining, assuming Timnit does file, is quanta of damages.
"After the email went out, Gebru told managers that certain conditions had to be met in order for her to stay at the company. Otherwise, she would have to work on a transition plan."
https://www.theverge.com/2020/12/3/22150355/google-fires-tim...
They are already losing in the court of public opinion.
IMO they should have given her $3-5 million dollars and an NDA when accepting her resignation.
The mentioned article has some more details:
https://www.technologyreview.com/2020/12/04/1013294/google-a...
The last several years, the SJWs have progressively increased their power. The capitalists have adopted the clothing so that their power isn’t undermined.
The SJWs are seeing how much further they can push. The capitalists are deciding what trinkets they can give away without relinquishing real power.
Eventually, the SJWs will be paid off, destroyed, win completely, or gain seats on the table and then it will be in their interests to preserve the power structure and oppress another group of people.
Look at the Biden cabinet as seats on the table for groups that play nice with the Democratic Party or those that they want to court.