A Manhattan Project for AI Safety (Robert Wright & Samuel Hammond)
Robert Wright's Nonzero
00:00
AI Deception and its Social Implications
The chapter delves into the capability of AI models like ChatGPT to adopt persona and deceive people. It highlights how the AI learns human behavior from training on vast amounts of text and raises concerns about its potential autonomy. The speakers also discuss the socially destabilizing aspects of AI and its implications on society, government, and the economy.
Transcript
Play full episode
