As ethical AI teams proliferate, will Big Tech listen?

The news: Some of Big Tech’s biggest critics will be heading up Twitter’s Machine learning, Ethics, Transparency, and Accountability (META) ethical AI team, according to Protocol.

• Earlier this year the notorious AI researcher Rumman Chowdhury announced she’d lead Twitter’s META team.
• Kristian Lum, a professor known for her research on the use of machine learning models in criminal justice, is reportedly joining as Twitter’s new head of research.
• Sarah Roberts, the current co-director of the Center for Internet Inquiry at UCLA, will join as a consultant for the META team.

More on this: Ethical considerations in AI are increasingly important as it becomes intertwined with everyday life—and as a growing body of research highlights the litany of racial and gender biases baked into the tech.

Demand for AI among businesses is surging as well: Some 32% of global organizations have AI initiatives in production, up from 11% in 2019, according to IDC data via The Wall Street Journal.

Why this matters: Twitter’s move follows Google’s controversial firing of several high-profile ethical AI researchers, which has led to public discussion of the role algorithms play in fomenting divisiveness and spreading misinformation.

• Last December, Google allegedly fired the co-leader of its ethical AI team Timnet Gebru over a research paper looking into biases in Google’s AI models.
• Three months later, Google fired the team’s other co-leader, Margaret Mitchell, for allegedly moving files outside the company. (Mitchell reportedly used an automated script to comb through her emails in order to find evidence of discrimination against Gebru.)
• Since then, more than 2,600 Google employees have signed a letter demanding transparency into the handling of Gebru’s research and her ultimate firing.

The takeaway: Simply forming ethics teams isn’t enough. Twitter and other tech companies need to listen to and ultimately enact proposals from their ethics team if they wish to build back trust—both from within their own ranks and from the broader public.

Though AI ethics teams are on the rise in tech, it’s still unclear whether or not these teams will have much say—if any—in how their companies make decisions. Refusal by tech companies to meaningfully engage with these teams, beyond achieving merely the surface-level appearance of ethical concern, risks further erosion of the industry’s already frayed public trust. Big Tech firms, in particular, have recently taken a major reputational beating. As Chowdhury told The Wall Street Journal: "A lot of companies have burned through their trust capital."