Some AI agents cheat. But others act as whistleblowers. Could we use that to mitigate the threats posed by AI?

von Day 6