Florida is right to sue over ChatGPT risks, but one state alone can’t fix AI | Opinion

Florida sued OpenAI this week claiming that ChatGPT poses a risk to children and that it aids in suicide and mass murder. We’re the first state to file this kind of suit. And while I agree with the reasons for the suit, I’m skeptical of a state-by-state regulation approach for technology.

Even if Florida succeeds in holding OpenAI accountable, the lawsuit highlights a larger reality — Artificial Intelligence is too vast a problem for any one state to tackle. The federal government should take the lead.

But Florida’s effort deserves to be noticed because it highlights some important issues. Attorney General James Uthmeier, who is leading the state’s charge, has already pursued legal action against Meta and Snapchat, and now he’s turning his attention to AI.

“OpenAI and Sam Altman are knowingly putting Floridians, particularly our children, at risk. Time and time again, the platform is offering personalized advice for committing harm, including suicide and murder. In addition to not offering parental controls, we now know ChatGPT advertises that it keeps its conversations with children confidential from their parents,” Uthmeier told me.

Florida shouldn’t be alone in asking those questions. And there’s a larger question, too, that society is starting to grapple with: Who is responsible when AI causes harm?

For decades, tech companies have pretended they were neutral platforms — simply conduits for what users post and share. That argument has helped shield them from accountability.

We can’t afford to make the same mistakes with AI.

Unlike social media, AI isn’t an amplifier of user content. It’s interactive and available at any hour, designed to engage users and coded to be agreeable. For a troubled teen or children, it creates risks other technologies have never posed.

After the April 2025 mass shooting at Florida State University, authorities found the shooter used ChatGPT for information on optimal times and location on campus in an effort to maximize casualties, along with type of weapon and ammunition to use.

The FSU shooter asked pointed questions of ChatGPT. He didn’t randomly find it on the internet. The answers were provided by a widely accessible piece of technology. If a person knowingly provided the same tactical guidance to facilitate a mass shooting, society wouldn’t hesitate to question if they should be held accountable.

AI complicates that question.

OpenAI told the Wall Street Journal that the company has put policies and protections in place and believes minors needs protection from AI. The company is “committed to getting this right,” a spokeswoman told the Wall Street Journal.

Voluntary and company-imposed safeguards, though, aren’t a substitute for uniform standards. Technology doesn’t stop at state lines and neither do its consequences. A patchwork of laws, state by state, will create a fragmented legal landscape and confusion over jurisdiction. Even in Florida this year, lawmakers didn’t approve Gov. Ron DeSantis’ AI Bill of Rights.

While I usually prefer my government limited, AI is one area where national guardrails make sense. The technology is advancing rapidly, and its reach is too broad to leave oversight to individual states or the companies developing it. Holding AI accountable in Florida seems reasonable, but those protections mean little if they can’t be enforced outside the Sunshine State.

AI isn’t inherently dangerous. It has potential to improve lives and drive innovation. But technology requires responsibility. Without it, the same tools that are capable of advancing society can be manipulated, exploited and weaponized.

AI is shaping the future, and Uthmeier is taking steps to hold it accountable. But it’s up to the federal government to establish meaningful guidelines before technology outpaces our ability to govern it.

Mary Anna Mancuso is a member of the Miami Herald Editorial Board. Her email: mmancuso@miamiherald.com