Skip to contents
An independent news site for, and about, the most powerful workforce on earth.
News

A Q&A with Dr. Hadas Kress-Gazit, robotics researcher who pulled-out of a Google workshop in support of AI ethicists

Dr Kress-Gazit talks with Techworker about how she made her decision to go public in her support of the ex-Googlers, and shared some advice for other workers considering the same.

Google’s recent firing of Drs. Timnit Gebru and Margaret Mitchell, and the company’s actions since, constitute one of the most severe transgressions the tech industry has seen in years. It has raised many questions, about the integrity of research published by the company, it’s treatment of Black employees, and the future of Google’s algorithms. 

Every time employees make headlines for speaking truth to power, fury ripples through the techworker community, adding ammunition to a brewing fight for better corporate policy and better worker protections. On an individual level, tech workers are left wondering, when is the right time for me to speak out too? 

These firings have, additionally, sent waves through the community of researchers and academic leadership in fields that overlap with Google’s technical interests and from whose pools Google often hires talent. A campaign titled #MakeAIEthical encourages non-Google employees to reject recruitment and decline Google research funding. 

Dr. Hadas Kress-Gazit took things a step further. The Cornell robotics professor not only pulled out of this week’s Google’s Machine Learning and Robot Safety Workshop, but told the company she was doing so in support of Gebru and Mitchell, and published the email she sent as a Twitter thread. Another Workshop invitee, Scott Niekum of the University of Texas at Austin, revealed that he too had dropped out in protest. Things went viral, and the fallout shifted into another gear.

Kress-Gazit spoke with Techworker on Wednesday about how she made her decision to go public in support of the ex-Googlers, and shared some advice for others considering the same. 

Q: What was your perception of Google before last December (a.k.a., before Dr. Timnit Gebru was fired)? 

A: I knew it was a big tech company, supporting research in areas I work in. I also knew it had diversity issues like many tech companies, especially in silicon valley. I know a few people who currently work or had worked there, but not many.

Q: When were you invited to the workshop? What was your perception of Google at that time? 

A: I was invited to the workshop on Jan 8, 2021. My perception was starting to change because of Dr. Gerbu’s firing and the gaslighting that happened after.

Q: How did you learn about Drs. Timnit Gebru and Margaret Mitchell’s firings? 

A: I learned through Twitter :). I do not know what is going on inside the company, but from the outside, the way she (and later Dr. Mitchell) was treated was textbook misogynoir.  

Q: What makes the time ripe right now to stop working with Google, rather than, for example, during the 2018 walkouts, or when the company lost their diversity chief in 2019? How did you know this was the best time to speak out? 

A: This is particularly egregious, but I guess my answer is more pragmatic — I did not have any interaction with Google, as a company, in the past several years. I gave a talk there in 2016 and I am involved with a conference that recieved Google support in the past (although I was not the contact or directly involved with the sponsorship), but other than that I had no interaction.  

Q: What practical calculations did you have to make before cancelling your participation in the workshop? Can you walk me through that thought process? 

A: My thought process was that I could not attend the workshop as if nothing had happened. I was debating for a while whether to say something at the workshop or just cancel my participation. I decided against attending and saying something because I felt that was not the right forum – it would be unfair to the organizers (who are researchers in my field) and to the other participants. 

As far as other considerations, as I wrote in my [Twitter] thread, this poses no risk for me. I do not have Google funding, I am tenured at a top university, my students have no problems finding jobs.

Q: Have you received any of the internet hate Dr. Gebru has posted about? And do you ever worry about getting funding from other large tech companies that might fear you’ll hold them accountable, too? 

A: No, I did not receive any hate/harassment (and hope it stays that way…). I also do not worry about not getting funding from other tech companies because I almost never apply for such funding. From her Twitter posts, it looks like Prof. [Emily] Bender [a University of Washington professor and co-author of the paper Dr. Gebru and Mitchell were fired for] also does not receive anywhere near the amount of harassment Dr. Gebru gets. I am saddened, angered, furious that a prominent outspoken, influential, Black woman is getting flooded with hate, but, unfortunately, I am not surprised. As I wrote in the thread – I admire her.

Q: If you could rewind a week, would you still send that email? And would you still post it as a Twitter thread, too? Are there any changes you would make?

A: Yes and yes. I would make one change, though. I would better delineate between Google, the company, which I have no interest in engaging with and that has behaved (externally) very poorly, and individual researchers, some of whom I know and highly respect. I would be clearer about my support for the researchers inside Google that are trying to create real change.

Q: If another researcher contemplated pulling out of a Google workshop, would you recommend they also be vocal about it? 

A: Not attending a workshop is really not a big deal. There are many workshops in many different venues, and people say no to workshops for a lot of different reasons. The answer is yes, but I do respect people’s different comfort levels. I think external pressure helps, and senior academics can help apply such pressure. But not everyone is comfortable with such actions.