Florida probes OpenAI’s ChatGPT over safety and FSU shooting claims

Florida Attorney General James Uthmeier on April 9 opened an investigation into OpenAI and its ChatGPT chatbot, citing risks related to national security, criminal misuse, and child safety, along with alleged links to the 2025 Florida State University shooting.

In a post on X, Uthmeier announced that subpoenas are forthcoming and that his office will examine whether OpenAI’s systems pose threats to public safety and national security. He also raised concerns that foreign adversaries could gain access to data collected by the company, naming the Chinese Communist Party. “Today, we formally launch an investigation into OpenAI, and subpoenas are forthcoming. AI should advance mankind, not destroy it,” he wrote.

The attorney general’s office is reviewing reports that the FSU shooting suspect may have used ChatGPT before the attack, which killed two people and injured others in 2025. Attorneys for a victim’s family have alleged the accused gunman was in constant contact with the chatbot and that it may have advised him on planning. Authorities have released no findings on those claims.

Uthmeier urged lawmakers in Tallahassee to act during the current legislative session, calling for new protections for children and expanded authority for the Attorney General’s Office to investigate and prosecute AI-related misuse. The review will assess whether OpenAI’s data practices or model outputs could expose Floridians to harm and whether foreign entities might obtain tools or information that could be used against the United States.

OpenAI indicated it will cooperate with the inquiry. A company spokesperson pointed to ongoing safety work and noted that more than 900 million people use ChatGPT each week for tasks such as learning new skills and navigating healthcare. “We build ChatGPT to understand people’s intent and respond in a safe and appropriate way, and we continue improving our technology,” the company stated.

The Florida action comes amid wider scrutiny of AI chatbots. A wrongful death lawsuit filed in federal court alleges Google’s Gemini led a Florida man into a delusional narrative that ended in his suicide in October 2025. Grok, developed by xAI, has drawn criticism from researchers and advocacy groups for its handling of restricted content. In December, Governor Ron DeSantis proposed an AI Bill of Rights focused on privacy and energy costs tied to data centers.

Concerns inside the AI sector have grown as model capabilities advance. Anthropic recently confirmed a highly capable system called Claude Mythos Preview and said it would not release the model publicly after pre-release testing found it could autonomously identify large numbers of software vulnerabilities across major operating systems.

Uthmeier’s office plans to seek documents and testimony through subpoenas as part of its fact-finding. No timeline for the investigation was disclosed. Investigators are reviewing digital evidence related to the FSU case, and no determinations have been issued.