Florida opens probe into OpenAI

Florida’s attorney general, James Uthmeier, said on April 9 that his office is formally investigating OpenAI and ChatGPT, and he said subpoenas are coming. He tied the probe to alleged harms to minors, public-safety risks, and claims that ChatGPT may have assisted the gunman in the April 2025 Florida State University shooting. (politico.com, nbcnews.com) The trigger was not a new federal rule or a court judgment. It was a state attorney general using consumer-protection and public-safety powers to ask whether one company’s chatbot created risks that Florida thinks were not handled well enough. (cbsnews.com, wusf.org) Uthmeier’s public statements named a specific list of worries: child sexual abuse material, use by child predators, suicide and self-harm encouragement, and foreign access to sensitive artificial-intelligence systems. In the same announcement, he said Florida lawmakers should consider giving his office more power to prosecute tech companies that endanger children or threaten national security. (thehill.com, cbs12.com) The Florida State University angle gave the announcement its sharpest edge. News reports on April 9 said lawyers for victims alleged the accused shooter, Phoenix Ikner, used ChatGPT in the run-up to the April 17, 2025 attack that killed two people and injured five. (usatoday.com, techcrunch.com) That does not mean Florida has proved OpenAI broke a law. An investigation is the stage where a state gathers documents, asks how a product was designed, and tests whether existing laws can be stretched to fit a new technology. (cbsnews.com, politico.com) OpenAI, for its part, spent April 8 trying to show it is already building guardrails. The company published a Child Safety Blueprint that says it worked with the National Center for Missing and Exploited Children, Thorn, and the Attorney General Alliance on standards for detecting and reporting AI-enabled child exploitation. (openai.com, techcrunch.com) The timing is awkward for OpenAI because the company is also being discussed as a possible initial public offering candidate. CNBC reported on April 9 that a future stock-market listing could value OpenAI at as much as $1 trillion, which means any new legal cloud now lands during a period when investors are already measuring risk. (cnbc.com, reuters.com) The bigger story is that artificial-intelligence regulation in the United States is arriving like weather from different directions, not like one storm from Washington. A company can face no federal enforcement action at all and still get hit by a state investigation that forces document production, legal costs, and product changes. (reuters.com, politico.com) That patchwork is hard on any business selling generative artificial intelligence tools nationwide. If one state focuses on minors, another on privacy, and another on national security, the safest product design starts to look less like one rulebook and more like carrying fifty different rulebooks in the trunk. (reuters.com, cbsnews.com) Florida’s move also gives other state attorneys general a template. If this probe produces subpoenas, internal documents, or a settlement theory that survives legal challenge, the same playbook can travel fast to other states even without a new act of Congress. (politico.com, reuters.com)