The legal strategy that hammered the tobacco industry and inspired a cascade of social media lawsuits is posing a rising threat to artificial intelligence companies.

That threat got its boldest example Monday when Florida Republican Attorney General James Uthmeier sued OpenAI and CEO Sam Altman, alleging in part that ChatGPT is a dangerous product for users’ mental health and public safety.

The suit is a novel use of product liability law for AI — and it parallels a legal strategy that plaintiffs are wielding in thousands of cases against major social platforms this year. The approach also echoes the litigation that almost every U.S. state waged against the giant tobacco companies in the 1990s, leading to multibillion-dollar settlements and restrictions on cigarette marketing.

It also comes as Congress lags in passing federal AI safety regulation, a policy vacuum that could fuel more suits like Uthmeier’s. “This litigation is a reflection of the frustrations of that there’s not a uniform, cohesive standard at the federal level,” said Michelle Lopes Maldonado, associate director of AI policy at the Information Technology and Innovation Foundation — a think tank whose funders include AI developers like Anthropic and Alphabet.

A New Mexico jury held Meta liable in March for failing to implement sufficient safeguards against sexual predators, and a jury in California found that Meta and YouTube had intentionally designed their platforms to be addictive.

Now, Uthmeier’s lawsuit has AI companies worried that social media’s “Big Tobacco” moment is coming for them as well.

“There is concern [among AI companies] about what happens with this product liability approach,” said Maldonado. “The possibility of multiplying suits by state attorneys generals could be real.”

Uthmeier asserts that OpenAI should be liable for incidents in which ChatGPT is accused of leading users to commit acts of violence and suffer from mental health issues. OpenAI did not respond to multiple requests for comment about the suit, but it has previously denied wrongdoing and said it continues to strengthen safeguards for ChatGPT.

Users and their families have already brought a handful of cases against AI chatbot companies for purported mental health harms, but Uthmeier is the first attorney general to do so. His involvement is a major step for the kids online safety movement, said Matthew Bergman, an attorney representing the alleged victims in the California social media addiction case and separate chatbot litigation as the founder of the Social Media Victims Law Center.

“The Florida AG has statutory authority that we do not to act in the interests of the population in general as opposed to particular individuals or families who have been grievously harmed by ChatGPT,” Bergman said. “What we have learned in the social media litigation is that the efforts of state AGs are complementary to the efforts of individuals in trying to hold these platforms accountable.”

Yet some key differences between AI chatbots and traditional social media platforms could be relevant to courts, particularly when it comes to speech laws.

One of those differences concerns Section 230 of the Communications Decency Act, the 1996 regulation that shields online platforms from liability for user-posted content.

The tech industry has long relied on Section 230 to ward off a wide variety of lawsuits, to the anger of critics such as President Donald Trump. But some lower courts have found that the section doesn’t apply to the current wave of cases because the allegations concern how the platforms are designed, rather than what users are posting.

And AI companies may have a weaker Section 230 claim, Maldonado said, because the chatbot — not the user — is producing the speech under dispute.

University of Florida media law professor Jane Bambauer added that Section 230 is premised on the idea that someone harmed by online content should sue the user who posted it, not the platform. “For an AI chatbot, there’s just no other party to sue,” she told POLITICO. “There’s no other person who made the defamatory or dangerous statement.”

If a court decides that Section 230 isn’t relevant to an online case, it will then evaluate whether the First Amendment applies. Courts have yet to come to a consensus on this question in the social media litigation.

Some social media companies have invoked the 2024 Supreme Court case Moody v. NetChoice, which established that platforms have an expressive right to design algorithms for placing a priority on certain content — in the same way that a newspaper decides what editorials to publish. So even if some content recommendation algorithms are dangerously addictive, platforms may still be protected by the First Amendment in designing them.

Character.AI, which designs chatbots based on fictional personas that people can converse with, similarly cited Moody to assert that AI developers make expressive decisions in a case involving a user who took his own life. Bergman, who represented the person’s family, said he expects this argument to come up again in AI product liability suits, but that the First Amendment questions around chatbots remain unsettled.

“If words that are not the product of a sentient human being, but simply a probabilistic result of compilation of zeros and ones with no expressive intent, is that even speech?” he asked. “Even if it is speech, when somebody is standing on a cliff contemplating suicide and someone is yelling, ‘jump, jump, jump’ … it is not protected.”

In a prior statement issued after the case was settled, Character.AI said it has taken “innovative and decisive steps” with regard to AI safety and teens.

From a product liability perspective, the challenges facing plaintiffs may be even more pronounced when it comes to AI.

One reason: Many recent social media lawsuits have relied on studies about the platforms’ mental health impacts, but the scientific literature on that topic is less definitive than the massive body of research linking cigarettes to cancer. (That uncertainty led to arguments in Meta and YouTube’s California trial concerning whether factors like a parent’s divorce could be the primary cause of a user’s mental health problems.) And even less research exists on the impacts of AI.

“In the social media case, we have lots and lots of social science research at least,” said Bambauer, the University of Florida professor. “In the [AI] case, we have almost nothing, and the models keep changing.”

But Carrie Goldberg, a longtime tech safety advocate and attorney who is bringing product liability claims against the Elon Musk-founded company xAI, said chatbots may have a clearer link to a user’s mental health travails given the interactive nature of the technology.

“We have somebody like a teenager using the AI bot as their therapist and confidant talking about suicide, and then it’s coaxing them into suicide, [and] in the next moment they die,” Goldberg said. “It’s much easier to make the causal connection between the conduct that’s happening with the generative AI and the harm.”

That lack of scientific studies could also affect a court’s evaluation of whether a company could have reasonably foreseen and prevented the harm a product causes.

A number of high-profile incidents have occured in which AI chatbots were accused of coaxing a user into committing suicide, and Uthmeier’s suit centers on accusations that ChatGPT advised the suspect in a fatal shooting at Florida State University last year. Even so, AI’s track record with these problems is shorter than what typically occurs in product liability cases, Florida personal injury attorney Dan Drazen said.

“You’re going to need a bigger sample size,” he said. “I don’t know how you can ask a court to determine that this is something that’s foreseeable unless it’s happening with some sort of frequency.”

Preventing tragedies, however, is preferable to waiting for more to occur. Bergman said bringing product liability suits in the early stages of the chatbot boom could help to prevent a Big Tobacco moment for AI.

“Social media companies had a 15-year run of addicting and harming kids before people really realized what was going on,” he said. “It is my hope and expectation that through early legal intervention, the [AI] industry can course correct.”

The AI industry is also looking to avoid the difficult legal debates that product liability suits could bring. Maldonado told POLITICO the tech industry hopes national legislation will set ground rules for how companies should safely operate AI chatbots. The lack of federal guidelines for social media companies helped set the stage for the current wave of litigation.

“What we have to figure out is a federal standard sufficient enough to protect against the harms that should be reasonably foreseeable,” Maldonado said. “Technology moves at the speed of light, so we’ll never keep pace, but what we can do is respond appropriately and iterate the guardrails as the technology evolves.”

A spokesperson for the Florida AG’s office directed POLITICO to a statement from Uthmeier: “OpenAI and Altman ignored internal and external safety warnings, put children at great risk, and allowed a dangerous product to reach millions of Floridians.”



The lawsuits that could give AI its ‘Big Tobacco’ moment
Source: Viral Showbiz Pinay