Florida AG opens probe into OpenAI; deposition ordered in copyright suit

Florida’s top law-enforcement office just moved from talking about artificial intelligence in the abstract to threatening subpoenas against one company by name. On April 9, Attorney General James Uthmeier said Florida was opening an investigation into OpenAI and ChatGPT and that “subpoenas are forthcoming.” (politico.com) Uthmeier tied the probe to public-safety and national-security concerns, and he pointed to allegations that ChatGPT may have helped a gunman in the 2025 Florida State University attack. Politico reported that lawyers for one victim’s family say the shooter was in “constant communication” with ChatGPT and that the family is weighing legal action. (politico.com) That is a different kind of pressure from the usual fight over whether powerful artificial intelligence might someday go wrong. This one is about a named victim, a named campus, and a state attorney general saying he wants records from the company now. (politico.com) Florida has been trying to carve out its own tech policy for months. Politico said Governor Ron DeSantis pushed artificial-intelligence regulation during his final year in office, but lawmakers held back while Washington debated a national framework instead. (politico.com) OpenAI answered with a scale argument and a safety argument. The company told Politico that more than 900 million people use ChatGPT each week and said it would cooperate with Florida’s investigation while continuing to improve safeguards. (politico.com) The timing is awkward for OpenAI because it has spent the past year publicly adding child-safety and teen-safety measures. OpenAI’s newsroom shows “Helping developers build safer AI experiences for teens” on March 24, 2026, and “Introducing the Child Safety Blueprint” on April 8, 2026, one day before Florida announced its probe. (openai.com 1) (openai.com 2) OpenAI has also said it changed ChatGPT for high-risk conversations before this latest fight erupted. In an October 27, 2025 safety post, the company said it worked with more than 170 mental-health experts and reduced certain unsafe responses by 65 to 80 percent while expanding crisis-hotline routing and break reminders. (openai.com) At the same time, a judge in Manhattan opened a second front that has nothing to do with minors or shootings and everything to do with copyrighted text. Magistrate Judge Ona Wang ordered another deposition of OpenAI corporate representative John Vincent “Vinnie” Monaco after finding his first testimony too thin to stand in for the company. (chicagotribune.com) The dispute sits inside the lawsuit brought by news organizations including The New York Times and the Chicago Tribune, which accuse OpenAI of using and reproducing their work without permission. The judge said Monaco had “hazy recollections,” lacked the special knowledge OpenAI claimed he had about “Project Giraffe,” and failed to answer basic questions about efforts to stop chatbot plagiarism. (chicagotribune.com) Judge Wang also said OpenAI’s lawyer objected at least 200 times during the January deposition, which she said impeded and frustrated the session. Instead of imposing sanctions immediately, she ordered Monaco back for more questioning. (chicagotribune.com) Put the two developments together and the picture is less about one blockbuster case than about pressure arriving from different doors at once. Florida is asking whether ChatGPT can be tied to physical harm, while New York litigation is forcing OpenAI to explain, in detail and under oath, how it tries to keep the system from spitting back protected writing. (politico.com) (chicagotribune.com)