As members of the Labor Tech Research Network, we share concerns about Timnit Gebru’s exit in December 2020 from Google, where she researched the ethics of artificial intelligence (AI; see go.nature.com/3ila). This has adverse implications for Black women working in tech companies, which are already known for their lack of diversity (D. Castelvecchi Nature 589, 12–13; 2021).
The research of Gebru and others has revealed the shortcomings of facial recognition for dark-skinned people, how automation can both replicate and obfuscate discrimination and how algorithmic systems reproduce racial exclusion and environmental racism. AI-ethics scholars cannot do their jobs properly if such conclusions could result in dismissal.
Given the potential for bias to creep into AI systems (see go.nature.com/3sxn), tech companies have a societal responsibility to retain critical voices. They should publicly affirm that all workers can pursue collective action, as is their right by law. And they should appoint independent ethics boards of seasoned researchers who can speak publicly on related issues and make formal recommendations to the governing board. To uphold the future of AI ethics, they should open the company to outside researchers and support underfunded tech-education programmes, including those at historically Black institutions.