The first move made by Google’s new artificial intelligence ethics lead? Firing her predecessor after she protested the company’s artificial intelligence ethics.
Margaret Mitchell, the former co-lead on the AI ethics team, tweeted on Friday evening that she’d been fired by the internet giant. The firing is meant to bring an end to a saga that began in December, when Mitchell’s co-lead, Timnit Gebru, was axed after a dispute over a research paper. Mitchell’s Google accounts were locked in early January after she was allegedly using scripts to find emails that displayed discriminatory treatment of Gebru, which she then allegedly leaked outside the company.
— MMitchell (@mmitchell_ai) February 19, 2021
Gebru is famed in AI circles for her work on algorithmic bias against people with dark skin. When the company asked her to retract a research paper that was critical of large-language models, which both write and evaluate text and are crucial to Google’s business, she refused to do so. She aired her frustrations with a number of colleagues over email, then told Google that she’d leave if it did not satisfy certain conditions. Google was uninterested in any terms and immediately fired her. Her termination drew outcries and protest both inside and outside Google.
Meg Mitchell, lead of the Ethical AI team has been fired. She got an email to her personal email. After locking her out for 5 weeks.
There are many words I can say right now. I'm glad to know that people don't fall for any of their bull.
To the VPs at google, I pity you.
— Timnit Gebru (@timnitGebru) February 19, 2021
It didn’t help that CEO Sundar Pichai promised to investigate the dismissal but was lukewarm on taking any actions that would shore up the integrity of the ethics team. That drew further outrage On Thursday, he introduced Dr. Marian Croak, a long-time Google employee, as the new head of the responsible AI ethics team. In a Q&A published on Google’s blog, she took a not-so-veiled swipe at the acrimony that had preceded her arrival.
“There’s a lot of dissension, a lot of conflict in terms of trying to standardize on normative definitions of these principles,” she wrote. “Whose definition of fairness, or safety, are we going to use? There’s quite a lot of conflict right now within the field, and it can be polarizing at times. And what I’d like to do is have people have the conversation in a more diplomatic way, perhaps, than we’re having it now, so we can truly advance this field.”