Jeff Ladish: When AI Models Know They're Being Tested They're Likely To Be On Their Best Behavior, And When They Don't Think They're Being Tested They're More Likely To Scheme Or Show Bad Behavior
AI Governance | How the AI Governance System Works | "For the First Time It's Possible to Completely Eliminate Privacy." - Yuval Noah Harari + Every Building Mapped, Every Movement Recorded, Every Behavior Scored
"Nudging" - Using AI to guide people's behavior in a positive direction - Marianna B. Ganapini is an associate philosophy professor at Union College (USA)