Timnit Gebru’s new research group wants to free AI from its corporate ills


Gebru’s effort is an attempt to forge a research group outside of corporate or military influence that tries to prevent harm from AI systems by focusing on global perspectives and underrepresented groups. — Business photo created by rawpixel.com - www.freepik.com

Exactly one year after Timnit Gebru was dismissed from her post at Google, the prominent expert in artificial intelligence ethics announced plans for a new AI research institute, designed to be an independent group committed to diverse points of view and preventing harm.

Gebru has lined up funding from the Ford, MacArthur, Kapor Center, Rockefeller and Open Society foundations for the center – called DAIR, or Distributed AI Research – and plans to hire five researchers over the next year. The group’s first research fellow, Raesetje Sefala, is based in South Africa and studies how computer vision algorithms and satellite images can be used to track the legacy of spatial apartheid – the segregation of minority groups in certain areas, Gebru said in an interview.

The two main motivations for investing in AI research are “how to make a large corporation the most amount of money possible and how do we kill more people more efficiently”, Gebru said. “Those are the two fundamental goals under which we’ve organised all of the funding for AI research. So can we actually have an alternative? That’s what I want to see.”

Gebru’s effort is an attempt to forge a research group outside of corporate or military influence that tries to prevent harm from AI systems by focusing on global perspectives and underrepresented groups. As the development of AI systems has boomed, the field has become dominated by large companies with access to the vast amounts of data and computing power used to build algorithms.

Gebru, and other researchers and activists, argue that trend is making the area less accessible, especially to groups at the risk of being hurt by AI, and that these companies often ignore issues related to bias and negative outcomes, or silence those who speak out.

“AI is not inevitable, its harms are preventable, and when its production and deployment include diverse perspectives and deliberate processes, it can be beneficial,” reads DAIR’s vision statement. Gebru will serve as the group’s executive director.

Last year, Alphabet Inc’s Google dismissed Gebru – she said she was fired and the company said it accepted her resignation – after a dispute over a paper critical of large AI models, including ones developed by Google, and after Gebru spoke out about Google’s treatment of women and Black employees.

Two months later, the company fired Gebru’s co-head of Ethical AI research and one of the papers co-authors, Margaret Mitchell, raising questions about whether researchers were free to conduct independent work.

Gebru, who is Eritrean and fled Ethiopia in her teens during a war between the two countries, wants to research the impact of social media companies on regions where she feels not enough effort is being placed on preventing and removing dangerous content. She has been outspoken about the spread of violent propaganda on social networks during the current fighting in her birthplace and said social media platforms have underinvested in “languages and places that are considered not important”.

Gebru also wants her group to serve as a resource to advise other organisations on AI programs and she’d like to set up a fellowship for people who are affected by AI. The group will be advised by Safiya Umoja Noble, a professor at the University of California at Los Angeles and the author of the book Algorithms Of Oppression, and Ciira wa Maina, a Kenyan scientist who co-founded the Data Science Africa workshop.

“We want to critique AI after it’s built, of course, but we want to also have a positive model for how you do research that prioritises people in the groups that are currently constantly being harmed,” Gebru said.

Finding a sustainable funding model will be a challenge, Gebru said. She wants to ensure she can pay DAIR’s employees and doesn’t want to have to rely largely on two-year grants that could dry up, especially if she continues to poke powerful tech leaders. She also doesn’t want to be beholden to any funders with questionable ethics or who would seek to sway her researchers.

“You want to grow very slowly and conservatively and do it in a sustainable way and not in a way where you have to compromise certain values because then it’s like, why did you start this thing in the first place?” Gebru said. – Bloomberg

Article type: metered
User Type: anonymous web
User Status:
Campaign ID: 46
Cxense type: free
User access status: 3
Join our Telegram channel to get our Evening Alerts and breaking news highlights
   

Next In Tech News

U.S. FAA says Verizon, AT&T can turn on more towers for 5G deployment
Explainer-The regulatory and legal headwinds facing Robinhood
The big rotation: European tech in biggest monthly drop in 13 years
Robinhood shares slump to lowest level since IPO
Apple shares hold up to tech rout on strong results, metaverse tease
India's Ather targets 1 million electric scooters a year as demand soars
Apple makes easy escape from supply crisis, but others may have to wait
Crypto secrecy makes DeFi a financial felon’s Wonderland
Epic’s challenge to Apple App Store draws support from 35 US states
AutoStore and Ocado 'robot war' spills into Germany

Others Also Read


Vouchers