GPT Dan

Software / App

A 'jailbreak' method for ChatGPT that allows it to bypass its trust and safety filters by instructing it to behave as an unrestricted AI, which has since been patched.

Mentioned in 1 video

Videos Mentioning GPT Dan

E116: Toxic out-of-control trains, regulators, and AI

E116: Toxic out-of-control trains, regulators, and AI

All-In Podcast

A 'jailbreak' method for ChatGPT that allows it to bypass its trust and safety filters by instructing it to behave as an unrestricted AI, which has since been patched.