Google’s AI ethics crew is asking executives to rebuild belief.
Google’s Ethical AI team sent Google executives a list of requests to “restore trust” after the removal of co-lead and renowned researcher Timnit Gebru.
Google’s Ethical AI team, which claims to advise on research, product, and guidelines, wrote a six-page letter to the company’s CEO Sundar Pichai, AI chief Jeff Dean, and technical vice president Megan Kacholia. The letter, titled “The Future of Ethical AI at Google Research,” watched by CNBC, lists the demands made by executives, including removing Kacholia from the group’s reporting structure, renouncing retaliation, and reinstating Gebru a higher level.
“In the past two weeks we have heard that Google leadership has pledged to continue the important work of diversity, equity and inclusion within Research PA and to support the Ethical AI team,” the letter said. “But we need to remember that words that are not paired with actions are signs of virtue. They are harmful, evasive, defensive, and show the inability of leadership to understand how our organization is part of the problem.”
Google did not immediately respond to requests for comments on the letter.
Correspondence with employees depicts distrust of leadership two weeks after Gebru, a well-known artificial intelligence researcher and technical co-lead of Google’s ethical AI team, tweeted that Google fired them because of a disagreement over a research paper , which examined the bias related to artificial intelligence. Gebru, who spoke openly about the company’s treatment of black employees, claimed the treatment showed a broader pattern on Google and sparked industry support, including a petition signed by thousands of Google employees and industry peers.
Last week, Sundar Pichai, CEO of Alphabet and Google, emailed employees apologizing for the distrust that Gebru had in the company and the industry when Gebru left.
Earlier Wednesday, several members of Democratic Congress, including Sens. Elizabeth Warren and Cory Booker, sent a letter to Pichai asking for details of his “review” of Gebru’s dismissal, as well as Google’s efforts to mitigate discriminatory prejudice in major language models to Jeff Dean first stated that this was a reason to reject the Gebru team’s paper. “The universe of elite AI researchers is limited in size, and Google’s staff disclosure guidelines have the potential to significantly influence academic and public discourse on AI,” the congressional letter said.
“Google’s short-sighted decision to fire and retaliate against a core member of the Ethical AI team makes it clear that we need swift and structural changes if this work is to continue and if the legitimacy of the entire field is to be maintained.” Letter from Google’s AI Ethics team reads.
The employee’s letter also asks for the removal of Kacholia, which, according to the team, informed Gebru about their removal without notifying Gebru’s direct manager. “We have lost confidence in Megan Kacholia and are demanding that she be removed from our reporting chain,” the letter said. “Samy Bengio, Timnit’s former manager, should report directly to Jeff Dean.”
The group asks Jeff Dean and Kacholia to apologize to Gebru and be “held accountable”. She explains that her dismissal has harmed Gebru and the Black Community on Google. “We see it as essential to remove Megan from our management chain.”
Details on Gebru’s shots are also requested. “Without understanding the process and the level of involvement of individual leaders, it is impossible to drive the structural changes needed to ensure this harmful behavior does not recur,” the letter said.
It also called for employees who defended Gebru not to be punished with “retaliation”.
“We know from Google’s retaliation in the past that they often harm and marginalize workers without firing them directly,” the staff explain. “Instead, the company will rank them badly in future performance review cycles, shifting their daily workload to something less desirable, and placing other direct and indirect limits on career growth.”
The staff also called for “explicit commitments that allow research to discuss the harms of certain technologies, including those that may be related to Google’s interests, products and research areas,” the letter reads, calling for a deadline for the end of the first quarter of 2021.
The staff also called for Gebru to be reinstated at a higher level.
“The removal of Timnit has had a demoralizing effect on our entire team,” the letter said. “She (Gebru) and Meg Mitchell had built a diverse, productive team that succeeded in a psychologically safe environment. Raising Timnit to their position at a higher level would go a long way toward restoring trust and rebuilding our team environment. “
Look now: The renowned AI researcher says Google abruptly fired her and sparked industry-wide criticism