[ad_1]
Focused on three groups in terms of increasing diversity in the year Artificial intelligence say they won’t fund it anymore Google. In a joint statement released Monday, Black in AI, Queer in AI and Widening NLP said they had protested against Google’s treatment of the former. Timnit Gebru, head of the AI ethics team and Margaret Mitchell, as well as former contractor April Christina Curley, a black woman who is black.
“The potential of AI technologies to cause special harm to members of our communities weighs heavily on our organizations,” the document says. “Google’s actions in recent months have caused tremendous damage to our entire community. Not only have they caused damage, they have set a dangerous precedent for finding out what kind of research, defense and revenge are allowed in our community.”
The document is endorsed by the groups calls made in March Google’s current and former employees are rejecting Google funding for academic conferences and officials are expected to establish stronger whistleblower protections for AI researchers.
This is the first time in the short history of each of the three organizations that they have been denied financial support by a sponsor.
Monday’s announcement marks a recent drop in allegations of interference in research into Google’s treatment of blacks and women and AI that will be published in academic conferences.
In March, the organizers of the Fairness, Accountability and Transparency (FAccT) conference denied Google grants and researcher Luke Stark for $ 60,000 in grants to Google. Luca Soldaini, Queer in AI organizer, told WIRED that the organization received $ 20,000 last year from Google; The NLP expansion received $ 15,000 from Google.
Cochair Xandra Schofield said the Widening NLP, created in 2017, was in line with the goal of bringing more women into the field to sign a joint statement because Google’s actions are inconsistent with the group’s role of supporting under-represented researchers. Mitchell was the co-founder of the organization. Haley Lepp added that by expanding NLP cochair, “by helping these scholars, we also want to help them with their research and the ability to conduct research that may be critical to the effects of AI.”
Membership groups such as Black in AI, Queer in AI, and Widening NLP are nonprofits that protect and replace people who have historically been underrepresented in the machine learning community. They operate separately from automatic learning conferences, but can attract hundreds of participants to workshops or social events organized at most of the conferences attended. In recent years, membership groups have been set up for people with disabilities and for Jews and Muslims.
The soldier said Queer in AI has come out in opposition to what non-trans and binary authors who want to update publications after Google Scholar changed their names.
“We’ve had bad or very bad experiences with that, and Google has had a very bad side,” he said. Requests to change Google’s names often go unanswered, he said.
He is the co-founder of Gebru Black’s AI. The paper discussed at the time he says he was released, about the risks that large language models pose to marginalized communities, was eventually published with Black identifying him as an author in AI. In a speech last week at the International Conference on Learning Delegations listed by Google as a platinum sponsor, Gebru declined to review articles submitted to machine learning conferences edited by lawyers.
“Academics don’t have to cover bets, they have to take a stand,” Gebru said. “This is not an intention. It’s about power, and multinational corporations have too much power and they need to be regulated. “
Black in AI, co-founder of Rediet Abebe, will be the first black member of the Berkeley Department of Electrical and Computer Engineering in California. he pledged not to take it last year Google’s money to reduce the company’s impact on AI research.
[ad_2]
Source link
