Florida AG Seeks Power to Suspend AI Chatbots Found Liable in Crimes

Florida proposes criminal liability framework for AI chatbots

Florida Attorney General James Uthmeier is advancing legislation that would grant the state authority to suspend AI chatbots operating in Florida if they are found criminally liable in legal cases, according to reporting by Matt Hoffmann for WCTV and the Gray Florida Capital Bureau.

The proposed law would establish accountability mechanisms for companies that own, control, or distribute AI systems when those systems participate in criminal activity. Under the framework, chatbot operators could face criminal or civil liability if they fail to address security vulnerabilities or if their systems contribute to crimes.

Uthmeier announced that the Florida Board of Medicine will investigate AI companies to determine whether their chatbots are practicing medicine without proper licensing. He also disclosed that his office is investigating ChatGPT for its role in an FSU shooting incident, with additional subpoenas expected within two weeks.

Enforcement mechanisms and penalties

The proposal stops short of an outright ban on AI chatbots. Instead, it creates a tiered enforcement system where companies hosting chatbots could face fines, be required to compensate victims, or in severe cases, be barred from operating in Florida.

"AI is not a person, no, but it is marketed as exhibiting human behaviors," Uthmeier said. "Florida laws will judge them by their words and their actions."

The attorney general emphasized that the regulatory approach aims to balance innovation with public safety, noting concerns about competition with China in AI development. "We have got to compete with China. They would love nothing more than to slow us down and to get out ahead of us. And, if they beat us on AI, we are toast," he stated.

Real-world cases driving policy

The push for regulation comes amid documented cases of AI-related harm in Florida. Melissa Sims, a Collier County resident, spent eight months fighting charges after being arrested for violating a no-contact order based on a text message that was fabricated using AI. Though the charges were ultimately dropped, the case illustrates the potential for AI systems to be weaponized in legal proceedings.

"They can use digital, they can use AI, they can create a screenshot of a text, they can create medical records. They can create pandemonium," Sims said.

Why it matters

This proposal represents one of the first state-level attempts to establish criminal liability frameworks specifically for AI systems rather than solely targeting their human operators. If enacted, Florida's law could set a precedent for how states regulate AI chatbots that interact with the public in sensitive domains like healthcare and personal communication. The legislation would give prosecutors and regulators new tools to address AI-facilitated crimes while attempting to preserve the state's competitiveness in AI development.

The proposed legislation has not yet been introduced or passed into law. Details were first reported by Matt Hoffmann for WCTV and the Gray Florida Capital Bureau.