Current initiatives to fix the issue are too narrow and shallow to be effective, according to a new report from the nonprofit AI Now Institute. Women only make up 10 and 15 per cent of AI research staff at Google and Facebook, respectively. Racial diversity is even worse—Black workers only make up 2.5 per cent of Google’s entire workforce, and four per cent of Facebook’s and Microsoft’s. (MIT Technology Review)

Talking point: The report says the focus on “women in tech” is too myopic, as it’s likely to privilege white women over others, as diversity initiatives tend to target gender without looking at intersections of other identities, like race. It also says that centring the school-to-industry pipeline via STEM education also doesn’t address workplace issues like harassment, unfair compensation and tokenization that disproportionately target women, trans people and people of colour, driving them out of the industry altogether. As artificial intelligence gets deployed in more industries—from hiring to housing to criminal justice—the implications of the report’s findings become more wide-ranging. The technology itself has already shown its creators’ biases, from devaluing women’s resumes to being unable to recognize darker-skinned faces.