Google CEO apologizes for handling of departure of AI expert
Pichai’s involvement shows how quickly the episode has spiraled out of control for Jeff Dean, Google’s head of AI and one of the CEO’s top deputies.
Google Chief Executive Officer Sundar Pichai apologized to employees on Wednesday and launched an investigation into how the company handled the departure of prominent AI researcher Timnit Gebru.
“I’ve heard the reaction to Dr. Gebru’s departure loud and clear,” Pichai said in a staff-wide email sent Wednesday that was reviewed by Bloomberg. “It seeded doubts and led some in our community to question their place at Google. I want to say how sorry I am for that, and I accept the responsibility of working to restore your trust.”
Gebru, a prominent artificial intelligence researcher best-known for showing how facial recognition algorithms are better at identifying White faces than Black and Brown ones, said she was fired last week after a dispute over an academic paper she co-authored calling out ethical issues related to technology that underpins important products like Google Search.
Pichai’s involvement shows how quickly the episode has spiraled out of control for Jeff Dean, Google’s head of AI and one of the CEO’s top deputies. The visceral reaction to Gebru’s departure -- and Dean’s initial response to it -- suggests discontent with management at one of the company’s most prized units.
The Alphabet Inc.-owned company has launched an internal review into how her departure was handled and the impact the incident has had on employees, especially those from under-represented minorities, the CEO wrote.
“We need to accept responsibility of the fact that a prominent Black, female leader with immense talent left Google unhappily,” Pichai said. “This loss has had a ripple effect through some of our least represented communities, who saw themselves and some of their experiences reflected in Dr. Gebru’s.”
The researcher wrote on Twitter that the CEO’s email didn’t go far enough.
The episode began when Google asked Gebru to retract the research paper, or at least remove the names of the Google employees involved, she said in a subsequent interview with Bloomberg. Gebru pushed back against the request.
Meanwhile, she cited the experience in an email group of company researchers and urged others to give up on a diversity report they were working on. “It doesn’t make a difference,” Gebru wrote in the email, which was obtained by Bloomberg. “There is no way more documents or more conversations will achieve anything.”
The incident spilled on to social media, spurring a petition signed by hundreds of current and former employees of Google, who said the real reason for her dismissal was her outspoken criticism of Google’s progress on improving conditions for people of color at the company.
Read more: Google’s Co-Head of Ethical AI Says She Was Fired by Email
The CEO’s message came after a heated meeting Tuesday with members of the Black Googler Network that left attendees questioning their trust in Google and their support of its AI work, according to screen shots of a moderated Q&A session that followed.
While Dean and Megan Kacholia, the vice president who notified Gebru of her termination, addressed the group, they didn’t take questions from attendees. Instead, they had a separate moderator field queries after they left, which angered some employees, many of whom spoke about quitting, according to one person familiar with the events. Other employees called the meeting “tone deaf” and questioned Google’s insistence to try to “sweep this wrongdoing under the rug,” according to messages reviewed by Bloomberg.
“We say we are committed to doing the right thing,” wrote one Black Googler who submitted a query during the session. “The right thing is to acknowledge that what happened to Timnit wasn’t right. I haven’t seen any of the Brain leadership acknowledge this. Without this it is hard for me as Black woman to trust Google.”
One questioner said Dean and Kacholia’s meeting seemed like a “form of gas-lighting” and asked what attendees were supposed to gain from it. Another asked leadership to “give one reason why I as a Black Googler in AI should support Google’s AI Fairness effort after Timnit’s disrespectful treatment.”