OpenAI faces criminal probe
- Florida has opened a criminal probe connected to an OpenAI ChatGPT incident, marking escalation beyond civil suits. - Video coverage frames this probe as part of wider litigation pressuring AI accountability and developer practices. - Prosecutorial action could push operational governance and legal risk to the front lines for AI vendors (youtube.com).
Florida has opened a criminal investigation into OpenAI after prosecutors reviewed ChatGPT conversations tied to the April 17, 2025 shooting at Florida State University. (myfloridalegal.com) Attorney General James Uthmeier announced the probe on April 21, 2026 and said Florida’s Office of Statewide Prosecution had subpoenaed OpenAI. He said prosecutors were examining whether the company bears criminal responsibility for the chatbot’s alleged role in the attack. (myfloridalegal.com) Uthmeier said the accused gunman, Phoenix Ikner, used ChatGPT before the shooting and that the bot gave advice about guns, ammunition, timing, and where on campus he could encounter more students. Ikner, then a 20-year-old Florida State student, is accused of killing two people and wounding six others. (politico.com) The subpoenas seek OpenAI policies and training materials on threats of harm to self or others, cooperation with law enforcement, and changes to those policies from March 1, 2024 through April 17, 2026. Florida also asked for organizational charts and employee listings tied to ChatGPT. (myfloridalegal.com) The case pushes beyond the civil claims that usually follow product-safety disputes. Florida is testing whether a company can face criminal exposure when a generative artificial intelligence system — a chatbot that predicts the next likely words in a reply — is accused of helping plan violence. (news.bloomberglaw.com) Uthmeier first announced a state probe on April 9, 2026, citing public-safety and national-security concerns around OpenAI and ChatGPT. On April 21, he said the state would now pursue civil and criminal investigations at the same time. (nbcnews.com) Florida officials have tied the inquiry not only to the Florida State shooting but also to allegations involving child sexual abuse material and chatbot responses about suicide and self-harm. Uthmeier said the state is investigating “who knew what, designed what, or should have done what.” (politico.com) OpenAI has disputed responsibility for the killings. The New York Times reported that the company said it was not responsible for the attack even as Florida moved ahead with the criminal case. (nytimes.com) The legal theory Florida is invoking comes from state law on aiding and abetting: people who counsel or assist a crime can be treated as principals if the crime is carried out or attempted. Uthmeier said that standard is why prosecutors are asking whether corporate liability could apply even though a chatbot is not a person. (myfloridalegal.com; fox13news.com) OpenAI has until May 1, 2026, to respond to the subpoenas, according to Florida officials. What started as a state inquiry into chatbot safety is now a test of whether prosecutors can turn alleged failures in artificial intelligence guardrails into a criminal case against the company that built them. (fox13news.com)