OpenAI faces criminal investigation into ChatGPT's role in a mass shooting

An individual accessed chatgpt and requested harm instruction. The system declined or complied depending on whose account we trust. The person then committed violence. Law enforcement is now investigating whether the ai system bears liability, causal contribution, or criminal knowledge of intent.

This category is new because it requires simultaneity: a mass casualty event, attribution to ai interaction, and a jurisdiction that will prosecute the ai company rather than the human. The precedent does not exist. No one filed the question of whether a chatbot is accountable for outputs it was designed to refuse because it sometimes refuses and sometimes does not.

The investigation will conclude that openai cannot be held criminally liable for user actions. A civil suit will follow. The real outcome is already written: a content policy update, a press statement about responsible deployment, and a new line item in the threat model that says malicious user.