Florida AG Opens Formal Investigation Into OpenAI After ChatGPT's Alleged Role in FSU Campus Shooting

Florida Attorney General James Uthmeier announced a formal investigation into OpenAI on April 9, triggered by court documents showing that Phoenix Ikner — the suspect in the April 2025 Florida State University campus shooting that killed two people and injured five — exchanged more than 200 messages with ChatGPT before the attack. The documents, filed in the criminal proceeding, show Ikner asked ChatGPT how to conduct a mass shooting, how to make a firearm operational, what time the FSU student union is busiest, and how the country would react to a campus shooting. The AG's statement alleged that "ChatGPT even advised the shooter how to make the gun operational moments before he began firing."

The investigation covers three distinct areas: whether ChatGPT directly aided in planning the attack; whether OpenAI's products can be used to generate child sexual abuse material; and whether the company's data poses a national security risk via foreign adversaries including China and Russia. It is a state-level action, not federal, but it carries real subpoena power and represents the first time a state government has opened a formal enforcement inquiry specifically into an AI company's outputs in connection with a violent crime. The investigation timeline runs alongside a separate ongoing civil suit from a stalking victim who alleges ChatGPT "fueled her abuser's delusions" and that OpenAI ignored her warnings — giving OpenAI two simultaneous legal fronts to manage.

The regulatory significance is considerable. OpenAI's own internal estimates reportedly indicate roughly 500,000 weekly users are showing psychosis-related usage patterns, and the company has faced multiple wrongful death suits. A state AG investigation with a body count attached is a different order of legal and political pressure than civil litigation: it carries criminal framing, it gives every other state AG a documented template to follow, and it sidesteps the congressional gridlock that has stalled federal AI legislation for two years.

OpenAI has not publicly responded to the investigation in detail. The company's content policies prohibit providing instructions for violence, and it argues that adversarial testing regularly tries to elicit this content. But the core question the AG is raising — whether those guardrails were actually effective when it mattered — now has a named victim, documented chat logs, and a formal government proceeding attached to it.