More than 1,600 Google employees have spoken out after a leading artificial intelligence researcher at the company claimed she lost her job for condemning the company’s diversity efforts and research integrity.
Another 2,500 academics, industry and civil society supporters have signed the Medium letter calling on Google to improve its commitment to research integrity after Dr Timnit Gebru, one of Google’s few Black female research scientists was dismissed earlier this month.
Gebru is known for her work on racist and sexist bias in AI technology, but according to the letter she faced consistent racism, research censorship, gaslighting and defensiveness before being fired.
Gebru led Google’s ethical AI division and had been working on a major paper on the ethical and environmental implications of large scale AI models.
According to the Medium letter, she and her colleagues had submitted the piece for internal review and in late November, five weeks after it had been reviewed and approved for publication, Google’s leadership censored it.
The paper was co-authored by Gebru, other Google researchers and external researchers and found that big tech companies needed to ensure AI systems designed to copy human writing and speech patterns did not entrench historical gender biases, according to Reuters.
Gebru on Wednesday said she had been fired for emailing an internal group of women and allies in the company’s AI unit to discuss Google’s censorship of the research paper and Google’s diversity programs.
“There is zero accountability. There is no incentive to hire 39% women:our life gets worse when you start advocating for underrepresented people, you start making the other leaders upset,” Gebru wrote in the email published by Platformer.
She said she felt “constantly dehumanized” at the company.
“There is no way more documents or more conversations will achieve anything.”
Gebru said she offered to pull her name from the paper provided the company detailed its objections and developed a stronger review process in future. If the company declined, she would arrange to resign.
I was fired by @JeffDean for my email to Brain women and Allies. My corp account has been cutoff. So I've been immediately fired :-)
— Timnit Gebru (@timnitGebru) December 3, 2020
Google said it could not accept her conditions and accepted her resignation but pulled it forward to be effective immediately, Gebru wrote on her Twitter, describing it as an effective termination.
Google maintains that Gebru resigned.
Google AI head Jeff Dean said Gebru’s paper “didn’t meet our bar for publication” in an email he shared publicly on Friday.
Thousands speak out
Prominent Google workers including Gebru’s co-team leader Margaret Mitchell have spoken out against the company’s decision, and by Monday more than 4,000 people had signed a Medium letter blasting the reasons for her departure.
“Dr Gebru asked them to explain this decision and to take accountability for it, and for their lacklustre stand on discriminatory and harassing workplace conditions,” the letter reads.
“The termination is an act of retaliation against Dr Gebru, and it heralds danger for people working for ethical and just AI — especially Black people and People of Color — across Google.”
The letter said Gebru was one of few people exerting internal pressure against the “unethical and undemocratic incursion of powerful and biased technologies into our daily lives”.
The letter, and its signatories, calls on Google to explain why the paper was rejected by leadership.
It also called on Google to make an unequivocal commitment to research integrity.
“We demand transparency to the broader public, including Google users and our colleagues in the academic community, about Google leadership’s decision to order Dr. Gebru and her colleagues to withdraw their research on large-scale language models,” the letter reads.
“This has become a matter of public concern, and there needs to be public accountability to ensure any trust in Google Research going forward.”