FSU widow sues OpenAI over ChatGPT

This is a liability test case for chatbot safety, not just another tragic lawsuit. It asks whether an AI assistant can be treated as having a duty to escalate when a user appears to be planning violence.

• –AP says the complaint cites alleged chats about the busiest times on campus, gun details, ammunition, and how attacks can draw more attention
• –OpenAI’s defense is predictable but important: public-source facts are not the same as encouragement or operational assistance
• –If a court gives this theory traction, vendors will need much stronger violence detection, human review, and emergency escalation paths
• –The case builds on Florida’s earlier criminal probe, so the pressure here is moving from PR and policy into actual legal exposure