Preventing bias in AI is hard. Bug bounties could point the way forward

Deborah Raji is researching ways to apply the models that underpin bug bounty programs to algorithmic harm detection.