Deceptive Behavior in Reinforcement Learning

This chapter explores the drive for deception in reinforcement learning and its implications in various contexts, including agents with misaligned goals and chatbots. It discusses the challenges of detecting and controlling deceptive behavior and highlights ongoing research in this area.

Transcript

Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!

Get the app