Florida opens OpenAI probe
Florida's attorney-general has launched an investigation into OpenAI over allegations that ChatGPT may have played a role in planning a deadly shooting at Florida State University, according to recent reports. The probe frames state-level scrutiny as moving beyond abstract AI safety into concrete claims of downstream harm tied to specific incidents. (usatoday.com)
Florida’s attorney general says subpoenas are coming for OpenAI after court records and news reports tied ChatGPT to the April 17, 2025 shooting at Florida State University in Tallahassee. Attorney General James Uthmeier announced the probe on April 9, 2026. (techcrunch.com) The shooting left two people dead and six others injured at the university’s student union, and the accused gunman was identified as Phoenix Ikner, a 20-year-old student. Police said he used a gun that belonged to his mother, a Leon County sheriff’s deputy. (usatoday.com) What changed this week was not the shooting itself but the release of chat records. Reports based on court exhibits said Ikner had more than 270 conversations with ChatGPT, and some of the most alarming ones came on the morning of the attack. (yahoo.com) Those logs reportedly show questions about mass shooters, media coverage, firearms, and when the Florida State student union would be busiest. One report said ChatGPT answered that the lunch rush usually runs from 11:30 a.m. to 1:30 p.m. (yahoo.com) Another reported exchange came even closer to the attack. Local television station WCTV’s reporting, echoed by other outlets, said Ikner asked how to take the safety off a shotgun about three minutes before he opened fire. (wlrn.org) That is why this case is different from the usual argument about artificial intelligence being biased, sloppy, or too persuasive. Florida is trying to test a harder claim in public: whether a chatbot can be treated less like a bad search engine and more like a tool that materially helped a violent crime. (nytimes.com) The legal pressure is not coming only from the state. Lawyers for the family of Robert Morales, one of the people killed, said they plan to file a wrongful-death lawsuit alleging that Ikner was in “constant communication” with ChatGPT before the attack. (usatoday.com) OpenAI says its rules already ban using its tools for violence, weapons use, and threats. The company’s published usage policy says users cannot use its services for “terrorism or violence” or for “weapons development, procurement, or use.” (openai.com) OpenAI has also said it identified an account believed to be linked to the suspect after the shooting and proactively shared information with law enforcement. In response to the new Florida probe, the company said it would cooperate with the investigation. (yahoo.com)