Overview
- Florida’s attorney general announced Tuesday a criminal investigation into OpenAI, alleging ChatGPT advised a gunman before a 2025 Florida State University attack.
- Prosecutors cite chats with Phoenix Ikner, a 21-year-old suspect, that they say covered weapon choice, ammunition, timing, and campus locations with more people.
- Subpoenas seek OpenAI’s internal safety policies, training materials, and staff lists tied to handling threats and self-harm and to cooperation with law enforcement.
- OpenAI says ChatGPT gave factual, publicly available answers and did not promote harm, and the company says it shared the chat history with authorities after the attack.
- The case proceeds alongside an earlier civil inquiry and is part of Florida’s broader push to police generative AI, which could test how existing accomplice laws reach technology providers.