Google’s Firing of Dr. Timnit Gebru by the Documents

Alison
3 min readDec 18, 2020
Screenshot of Dr. Timnit Gebru’s Twitter bio

Dr. Timnit Gebru, a black woman doing foundational and beyond-words-important research in AI Ethics, was fired by Google. I’m a white woman, and this is my summary of watching the developments on Twitter in the first few days. Originally written 12/4/2020.

I #BelieveBlackWomen. I especially know that when such an accomplished and respected black woman has the courage to share publicly what’s happened to her, and people still don’t believe her despite her reputation, that it is more extremely important than ever to support her wholeheartedly. So to me when she and the people on her team and the people who have experienced similar things with Google make statements then I personally believe it to be the truth and I will characterize it and refer to it as such here.

What I know to be true is that there were a complicated combination of circumstances and Google followed it’s abusive, probably illegal, patterns. The core documents involved are: a) an email Dr. Gebru wrote to the Google Brain Women and Allies group, b) a research paper Dr. Gebru co-wrote, c) feedback [I think it was written] Dr. Gebru received about her research paper, and d) an email Dr. Gebru wrote to Google giving two conditions she wanted them to meet otherwise she would quit.

For context, Dr. Gebru’s email to Brain Women and Allies is :fire:. It is one of the best things I’ve read all year, something I’ll reference for years to come — just the honest truth as I’ve painfully learned in the 20+ years I’ve been writing code. To summarize, she said stop doing the diversity work at Google (“writing documents”) because it hasn’t and won’t make a difference but, if you really want to do something, do work to make leaders be held accountable.

Let’s go to the research paper(b). It had been “pubapprove “approved”” by Google which means approved. Also, Google has a history of not punishing people who don’t go through the internal research review process. The process, is by the way, designed to keep corporate secrets not to do peer review. Then she wrote the email to Brain Women and Allies(b) after which she received anonymous feedback routed to her through the paper review team(c) about problems with citations in her co-authored paper(b). She wanted to know who was giving that feedback (secret feedback trick), she wanted to address them. They wouldn’t tell her who wrote them. She wrote an email to Google(d) saying she’d take her name off the research paper if they could meet 2 conditions, otherwise she would work with them on a last day at Google. Condition 1 was that they tell her who was newly critiquing her paper. Condition 2 wasn’t publicly known as of the last time I checked.

Which brings us to the email Google sent Dr. Gebru where they claim she resigned and that they had decided to make it effective immediately. I say Google because her boss didn’t know she’d been fired. They specifically mention in their email(d) that her email to Brain Women and Allies(b) “is behavior that is inconsistent with the expectations of a Google manager.” Talking about this as an issue with her paper(b) is just disingenuous [as Jeff Dean, SVP of Research at Google tried to/is trying? to do]. It’s also a double standard compared to how Google treats white men as this Googler notes.

To see and sign the petition standing with Dr. Gebru, see the Medium post. As of 12/17/2020 there were “2695 Googlers and 4302 academic, industry, and civil society supporters” signers. Please become one of them.

There’s also an FAQ follow-up and Dr. Gebru’s twitter account.

--

--

Alison

All Things Data and Databases. Knitter. I listen to #womenintech. She/Her/Hers.