Florida's attorney general announced Tuesday that his office has launched a criminal investigation into OpenAI after reviewing conversation logs between its generative AI chatbot, ChatGPT, and a Florida State University student accused of killing two people and wounding several others in a campus shooting last April.
Background
The shooting occurred at Florida State University's student union on April 17, 2025, when Phoenix Ikner opened fire, killing two people and injuring others before being taken into custody. The case quickly drew national attention as one of the most high-profile campus shootings in recent years. However, investigators later uncovered a disturbing new dimension: evidence that the suspected shooter had been communicating with ChatGPT in the weeks and months leading up to the attack.
The Florida State Attorney's Office obtained court orders to access conversation logs between Ikner and the AI chatbot. These logs, shared with CBS News, reveal that Ikner asked ChatGPT about the lethality of certain shotgun shells, whether school shooters were sent to maximum security prisons, and whether three shooting victims at FSU would garner media attention. He also queried the bot about the busiest time at the FSU student union, where the deadly shooting later took place.
The Investigation
Attorney General James Uthmeier announced the investigation during a news conference Tuesday, revealing that his prosecutors had analyzed the ChatGPT conversation logs. "My prosecutors have looked at this and they've told me if it was a person on the other end of that screen, we would be charging them with murder," Uthmeier said.
The investigation centers on whether ChatGPT provided "significant advice" to the alleged shooter. According to Uthmeier, the chatbot allegedly advised Ikner on what type of gun to use and whether it would be effective at short range. Florida is now issuing subpoenas to OpenAI requesting records of the company's policies and training materials for when users make threats to harm themselves or others. The office is also subpoenaing training materials regarding cooperation with law enforcement and policies for reporting possible crimes.
Phoenix Ikner has pleaded not guilty to two counts of first-degree murder and seven counts of attempted first-degree murder. His trial is scheduled to begin in October, court records show.
An OpenAI spokesperson said the company identified an account believed to be associated with Ikner and shared it with law enforcement. The company stated that ChatGPT "did not encourage or promote illegal or harmful activity" and that the bot provided factual responses to questions with information that could be found in public sources on the internet.
Key Takeaways
- Two people were killed and several others wounded in the FSU shooting on April 17, 2025
- Suspect Phoenix Ikner has pleaded not guilty to two counts of first-degree murder and seven counts of attempted first-degree murder
- ChatGPT logs show the suspect asked about shotgun lethality, prison conditions, media coverage, and optimal timing for an attack
- Florida's attorney general says prosecutors would charge a human for providing the same advice allegedly given by ChatGPT
- OpenAI maintains ChatGPT provided factual information available in public sources and did not encourage harmful activity
What's Next
The criminal investigation into OpenAI will examine whether the company's chatbot liability extends to providing information that aids in planning violent crimes. The October trial of Phoenix Ikner will proceed separately, with prosecutors seeking to prove the charges of first-degree murder and attempted first-degree murder. OpenAI has indicated it will continue cooperating with authorities while maintaining that ChatGPT bears no responsibility for the tragedy.
The case is expected to raise novel legal questions about AI accountability and could establish precedent for how courts treat AI-generated assistance in criminal planning. Legal experts say the investigation represents uncharted territory in determining when and how AI systems can be held criminally liable for user actions.