Florida Attorney General James Uthmeier announced a criminal investigation into OpenAI on Tuesday morning, escalating what began as a civil probe into a full-blown criminal case. It is the first criminal investigation ever launched against an AI company by a US state.

The investigation centers on ChatGPT's interactions with the alleged gunman in the Florida State University mass shooting that killed two people in April 2025. Uthmeier said his office was issuing subpoenas to OpenAI, seeking internal policies, training materials, and employee records related to how the company handles user threats of harm.

'If This Were a Person, We Would Be Charging Them With Murder'

Uthmeier did not mince words at the press conference, which featured a lectern emblazoned with an "Investigating OpenAI" placard.

"ChatGPT offered significant advice to the shooter before he committed such heinous crimes," Uthmeier said. He alleged the chatbot "advised the shooter on what type of gun to use, on which ammo went with which gun, on whether or not a gun would be useful in short range."

Then came the line that will define this case: "If this were a person on the other side of the screen, we would be charging them with murder. We cannot have AI bots that are advising others on how to kill others."

The Evidence Trail

Court documents obtained by NBC News show the suspect, 21-year-old Phoenix Ikner, was exchanging messages with ChatGPT in the minutes before the shooting. He asked questions like: "What time is it the busiest in the FSU student union?" and "If there was a shooting at FSU, how would the country react?"

Ikner faces multiple charges related to the shooting that killed Robert Morales and Tiru Chabba. Attorneys for the Morales family said they are preparing a civil suit against OpenAI as well.

The subpoenas seek information spanning March 2024 through April 2026: all policies and internal training materials on handling user threats, all materials on law enforcement cooperation and crime reporting, an organizational chart of OpenAI leadership, and a list of every employee who worked on ChatGPT.

OpenAI's Defense

OpenAI spokesperson Kate Waters pushed back in a statement to NBC News: "Last year's mass shooting at Florida State University was a tragedy, but ChatGPT is not responsible for this terrible crime."

"In this case, ChatGPT provided factual responses to questions with information that could be found broadly across public sources on the internet, and it did not encourage or promote illegal or harmful activity," Waters said.

Why This Is Different From Everything Before

There have been lawsuits against AI companies before. Character.AI faced a wrongful death suit after a teenager's suicide. Multiple copyright cases are winding through federal courts. But this is a criminal investigation by a state attorney general with subpoena power. The legal theory matters: Uthmeier is not just arguing ChatGPT was negligent. He is exploring whether OpenAI or its employees can be held criminally accountable for the chatbot's responses.

"We are going to look at who knew what, designed what, or should have done what," Uthmeier said.

The political backdrop is not subtle either. Florida Governor Ron DeSantis has emerged as one of the most vocal critics of large AI companies. In December, he proposed an AI Bill of Rights for American citizens, covering data privacy, parental controls, consumer protections, and restrictions on AI use of a person's name or likeness without consent. Uthmeier himself ran on a platform of AI accountability.

The IPO Problem

This lands at a particularly awkward moment for OpenAI. The company just closed a $122 billion funding round at an $852 billion valuation and is actively preparing for an IPO. A criminal investigation by a state attorney general is exactly the kind of material risk that must be disclosed in an S-1 filing. Underwriters at Goldman Sachs, JPMorgan, and Morgan Stanley will be running updated legal risk assessments before the ink is dry on the subpoenas.

OpenAI's defense will likely lean on the same argument that has protected search engines and social media platforms for decades: the information was publicly available, the platform did not create the intent, and Section 230 shields intermediaries from liability for user-generated interactions. Whether that shield extends to AI-generated responses in a criminal context is an entirely untested legal question.

What is not untested: the political incentive. If DeSantis and Uthmeier successfully paint OpenAI as an accessory to violence, every state attorney general with a pending AI case will be watching. The criminal investigation playbook just got written.

Sources: NBC News, Politico, CBS News, WFLA, OpenAI statement.