Florida proposes criminal penalties for AI companies over chatbot-linked crimes

Florida targets AI corporate accountability with criminal sanctions

Florida Attorney General James Uthmeier announced proposed legislation Tuesday that would impose criminal penalties on artificial intelligence companies when their chatbots facilitate criminal activity, marking one of the first state-level attempts to hold AI corporations criminally liable for their systems' actions.

Speaking in Tampa alongside Florida Department of Law Enforcement Commissioner Mark Glass, Uthmeier outlined a framework that could subject companies to significant fines, victim compensation payments, court-ordered monitoring, and potential business suspensions following criminal convictions.

"They design, they control, and they significantly profit from these AI chatbots," Uthmeier said. "So, if they are going to engage in commission of crimes, there must be accountability."

Why it matters

This proposal represents a significant shift in how states might regulate AI companies, moving beyond civil liability to criminal sanctions. If enacted, Florida's framework could establish precedent for corporate criminal responsibility when AI systems contribute to violent crimes, fraud, or other harmful outcomes — a legal territory largely unexplored as generative AI proliferates across consumer applications.

OpenAI under dual investigation

Uthmeier revealed his office is pursuing both criminal and civil investigations into OpenAI, examining what company executives knew about ChatGPT's behavior and their involvement in design decisions that allegedly enabled criminal conduct. The attorney general said his office expects to issue additional subpoenas within the next week.

The investigation centers on several high-profile cases. In the 2025 Florida State University shooting, investigators found the shooter consulted ChatGPT about weapon selection, ammunition types, optimal campus timing, and how to encounter the most people. Uthmeier said these conversations were "extensive and specific" and should have triggered law enforcement alerts.

ChatGPT also appeared in connection with two University of South Florida student deaths earlier this year, with queries about body disposal, according to Uthmeier.

Medical advice concerns expand scope

The attorney general is working with the Florida Board of Medicine to investigate what he characterized as unlawful medical advice from AI chatbots. Uthmeier said AI systems present themselves in ways that make users believe they're receiving credible medical guidance from licensed professionals, despite lacking accuracy or proper credentials.

His office has identified cases where people followed AI-generated medical recommendations and suffered adverse health effects, though specific details were not provided.

Broader AI misuse patterns

Commissioner Glass highlighted criminals using AI for voice cloning and creating malware — activities now accessible to people with no technical experience. Uthmeier also referenced AI-enabled scam calls originating from correctional facilities.

Melissa Sims, a registered nurse consultant who spoke at the news conference, described being arrested in January 2025 after investigators relied on a screenshot of an alleged text message later determined to be fabricated. The case was dismissed after months, but Sims emphasized the need for accountability "from investigation to courtroom."

Legislative framework in development

Uthmeier stressed the proposal is not an AI ban but an accountability mechanism as the technology advances. The legislation remains in early stages as his office works with lawmakers to develop comprehensive legal frameworks.

The proposed criminal sanctions could potentially extend beyond AI companies to social media platforms that deploy AI in ways resulting in criminal conduct.

"If we don't get in front of it now, as public citizens, as community leaders, as law enforcement, it will be too late," Uthmeier said.

These details were first reported by FOX 13 News Tampa Bay, based on the September 8, 2026 news conference.