WASHINGTON – This month, Florida became the first state to sue artificial intelligence behemoth OpenAI.

The lawsuit alleges, among other claims, that a 20-year-old gunman sought advice from ChatGPT before killing two people and wounding others at Florida State University last year. The suit, which seeks to hold OpenAI partly responsible for the mass shooting and other safety lapses that resulted in harms to Floridians, reflects society’s broader struggle to hold companies accountable as AI chatbot technology becomes increasingly ubiquitous in American life.

Data from a Common Sense Media survey conducted last year shows that 72% of teens have used AI companions at least once, with one in three teens reporting that they’ve turned to chatbots for social interactions and emotional support.

Families have sued AI companies over the suicides of loved ones they claim were linked to chatbot interactions. In one of the first high-profile cases, a 14-year-old took his own life in 2024 after developing a relationship with a Character.ai chatbot that his family said encouraged him to go through with the act. Character.ai agreed to a settlement with the family.

A spokesperson for Character.ai said in an email to Tech Policy Press that the company removed its open-ended chat feature for users younger than 18 last year.

As lawsuits accusing AI companies of contributing to wrongful deaths continue to emerge across the country, Congress has been working to draft and enact legislation aimed at protecting minors from potential harms posed by AI chatbots.

In April, the Senate Judiciary Committee unanimously voted to advance the GUARD Act, a bipartisan bill introduced by Senators Josh Hawley, R-Mo., and Richard Blumenthal, D-Conn. The legislation would impose broad protections for young people using AI chatbots, including age-verification requirements.

“There’s nothing missing. We have a framework for regulations,” Blumenthal told Tech Policy Press. “Hawley and I provide oversight and make sure that these models are safe and effective. But even more important is to prevent the chatbots from exposure to children and to impose age verification.”

If passed, the bill would ban AI companions designed to simulate relationships with minors, require chatbots to disclose their non-human status, and impose criminal penalties on companies that make chatbots that produce sexually explicit content available to minors.

“These are pretty basic things,” Hawley told Tech Policy Press. AI companies “don’t want to lose profit. That’s what they care about the most. What I care about most is protecting kids, protecting people, and protecting workers. We’ve got to put people over money. It’s really not that hard.”

States are taking action

The GUARD Act is awaiting consideration by the full Senate, though no date has been scheduled yet. A companion bill was also introduced in the House.

Congress has also introduced several other bills this year related to AI protections for minors, each with a different approach.

The CHATBOT Act includes requirements that companies provide parental controls for children’s AI use. The Youth AI Privacy Act would prevent companies from using minors’ personal data to train their AI models. And the KIDS Act package includes chatbot-specific measures that, for example, would prevent chatbots from claiming to be licensed professionals.

But in the absence of concrete federal progress, states are moving to fill in the gaps.

Since 2025, lawmakers in 49 states and the District of Columbia have introduced 464 bills related to chatbot safeguards and AI usage in health care, according to the National Conference of State Legislatures Artificial Intelligence Legislation Database. High-profile lawsuits over children's deaths linked to AI companions helped to spur this legislative wave. More than half of those states have successfully enacted at least one law.

California was the first state to pass a law regulating AI companion chatbots with protections specifically for minors. SB 243, introduced by California State Senator Steve Padilla, was signed into law by Governor Gavin Newsom last year and took effect on January 1, 2026.

Like the GUARD Act, the law requires chatbots to disclose that they are not human. It also requires chatbot providers to recognize signs that users may be expressing harmful thoughts, like suicidal ideation, and direct them to appropriate resources. In addition, it requires companies to filter out sexually explicit content for underage users.

“I think AI is the most substantial advent since the Industrial Revolution itself, which took a century to unfold, but here we are at breakneck speed, so we have an obligation to make sure that we're getting this right,” Padilla said in an interview with Tech Policy Press. “I think we can have some of the best tech, but we should also have the safest products, and we should protect folks and make sure that we're getting ahead of it before it's too late.”

Increased protections for minors have been a legislative priority for Padilla, who said the issue hit especially close to home following the death of Adam Raine, a 16-year-old Californian whose parents allege ChatGPT told him to end his life last year.

Padilla urged the federal government to take the issue more seriously and acknowledged his state’s role in informing the conversation around regulation.

“As the most populous, largest state in the union, we certainly have an impact that's appropriate given the size of the state,” Padilla said. “We have an opportunity, maybe even an obligation, and we can set standards that can impact what's going on throughout the rest of the country.”

Washington State Representative Lisa Callan shared a similar sentiment, noting that her state was taking action while Congress stalled.

“It's a high-risk ball game, and lives are being lost, so we have to act,” she said. “It's better if we can do it at a national level so we have consistency across the nation, but we can't wait, so Washington needed to take a step.”

Callan sponsored Washington’s HB 2225, which included similar provisions to California’s SB 243 by requiring chatbots to disclose that they are AI and prohibiting them from generating sexually explicit or harmful content.

She said she hoped that states could play a role in navigating a path to federal law.

“I'm concerned about preemption if the federal government weakens what we've already gotten done,” she said. “I think anything that we've done at each of the state levels that is stronger will help us inform what needs to happen and what's working.”

The path to federal regulation

Since returning to office for his second term, President Trump has taken a strongly pro-AI stance, emphasizing deregulation and private-sector innovation as key to maintaining US economic competitiveness.

In March, the White House unveiled a six-point national legislative framework addressing several prominent concerns around AI, including protections for minors.

The document called on Congress to give parents tools to manage their children’s online activity. It also endorsed guardrails similar to those proposed in the GUARD Act that shield minors from sexually explicit content or content that encourages self-harm.

Gaia Bernstein, a Seton Hall law professor specializing in tech policy, argued that AI should be regulated as a public health issue. She proposed that AI companies should be required to issue recalls for their products and ensure they are safe before releasing them to the public

Still, she acknowledged that such recall measures would be difficult to enact, but are necessary.

“There is the idea that this technology is so useful that we should never take it off the market, but I think we should start placing the risk of error the other way,” she said. “We want to make sure we don't cause harm with these technologies, even at the expense of some time off the market.”

Conversely, other policy experts argued that the GUARD Act’s provisions were too stringent and risked impeding minors’ access to this emerging technology.

“We all do agree that children should be protected online. The question, particularly with the GUARD Act, is whether it is the right tool,” said Aden Hizkias, associate policy director at the Chamber of Progress, a trade group that opposed the GUARD Act.

In a letter to Congress, Hizkias claimed that the definition of an “AI chatbot” was too broad and could hinder young users from accessing low-risk and educational tools like AI tutors and language-learning bots.

“It's a balance between parents having tools and ensuring that there is some form of control on what a kid can access, but not so much so that a kid just doesn't have access at all,” she said.