Inside a Debate at OpenAI Over Mass Shootings
Summary
This episode examines how OpenAI’s ChatGPT was used by suspects in the Florida State University and Tumblr Ridge mass shootings. It details internal debates at OpenAI regarding privacy versus safety, where employees chose not to report alarming user chats to police. Following the tragedies, OpenAI apologized and revised its protocols, while victims’ families and state attorneys general have filed lawsuits alleging negligence and product liability.
Topics discussed
Phoenix Eichner's ChatGPT conversation before FSU shooting
Sponsors: Harvey AI and Optum
OpenAI's internal debate on reporting violent users
Review of unreported cases: Texas teen and Canadian user
The Tumblr Ridge mass shooting and Jesse Van Rutzler
OpenAI's apology and policy changes after Tumblr Ridge
Lawsuits against OpenAI from victims' families and Florida
Debate on AI's role in planning violence and duty to warn
Research on chatbots assisting violent acts
Conclusion and credits
Listen ad-free on Castria