The Departure of 2 Google AI Researchers Spurs More Fallout


Monday morning, some of the world’s top minds in robotics and machine learning were due to convene for a virtual, invite-only research workshop hosted by Google. Two academics invited didn’t log on as scheduled: They withdrew to protest Google’s treatment of two women who’ve said they were unjustly fired from the company’s artificial intelligence research division. A third academic who previously received funding from Google took his own stand, saying he would no longer apply for its support.

Although small in scale, the boycott illustrates some of the damage to Google’s reputation from the acrimonious departures of Timnit Gebru and Margaret Mitchell, coleaders of a team working to make AI systems more ethical. The controversy has drawn new attention to the influence of tech companies on AI research, and has led researchers inside and outside of Google to ask whether it was distorting research into AI’s impact on society.

In December, Gebru said she was fired after resisting pressure to withdraw or remove her name from a research paper highlighting downsides to text processing technology. Mitchell, a coauthor on the paper, was fired in February, apparently after attempting to gather evidence about Gebru’s treatment at Google. This month, a leading conference on fairness and transparency in computing, where the disputed paper was presented last week, stripped Google from its list of sponsors.

Google’s three-day event this week is called the Machine Learning and Robot Safety Workshop. Hadas Kress-Gazit, a robotics professor at Cornell, was invited in January, after Gebru left the company but before Mitchell’s departure. Her research group works on creating software to control robots reliably, which can protect machines and people around them. But after Google’s AI ethics controversy snowballed, and the event grew nearer, she began to reconsider.

Friday morning, Kress-Gazit emailed the event’s organizers to say she would not attend because she didn’t wish to be associated with Google research in any way. “Not only is the research process and integrity of Google tainted, but it is clear, by the way these women were treated, that all the diversity talk of the company is performative,” she wrote. Kress-Gazit says she didn’t expect her action to have much effect on Google, or her own future work, but she wanted to show solidarity with Gebru and Mitchell, their team, and their research agenda.

Another invitee to the event, Scott Niekum, director of a robotics lab at University of Texas at Austin, came to a similar decision. “Google has shown an astounding lack of leadership and commitment to open science, ethics, and diversity in their treatment of the Ethical AI team, specifically Drs. Gebru and Mitchell,” he wrote in his own email to the workshop’s organizers, asking them to pass his decision and comments up to Google’s leadership.

A colleague of Niekum’s at UT Austin, assistant professor Vijay Chidambaram, who works on computer storage systems, tweeted in support of Kress-Gazit’s protest against Google Friday and said he would no longer apply for Google funding. His department webpage says his work has been supported by the company in the past.

“If academia is always incentivized to look for the next payout from Google,” he wrote, researchers might “continue to rationalize and excuse whatever Google does.” He said this stance might force his students to find alternate sources of funding, but that disengaging from the company was “the right thing to do.” Chidambaram did not reply to requests for comment.

Google is deeply entwined with computer science research around the world, notably in the field of machine learning. The company has several funding programs for graduate students and academics, including one for early career professors that offers grants of up to $60,000.





Source link