Tom Barnes answers Strongest real-world examples supporting AI risk claims?

[Tom Barnes]

For deception (not deceptive alignment) - AI Deception: A Survey of Examples, Risks, and Potential Solutions (section 2)