Eight states sued OpenAI on Wednesday, accusing it of misleading parents about ChatGPT’s security.
The coordinated lawsuits – filed individually in Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana – target the AI giant for the harm its generative AI chatbot allegedly caused to children in those states.
“OpenAI, they wanted to be first to market. They advanced ChatGPT at high speed, even though they knew they were harming children, lying to parents and breaking our laws,” Iowa Attorney General Brenna Bird said at a news conference announcing the lawsuit. “Everyone must respect the law.”
The company’s flagship AI model, ChatGPT, kicked off the generative AI boom in November 2023, when it exploded into mainstream culture, and quickly became the world’s largest chatbot. This summer, it surpassed one billion global monthly active users of the app.
In recent years, users have increasingly turned to it and other chatbots for everything from friendship to therapy to a host of agent tasks in their personal and professional lives.
Bird was joined at Wednesday’s news conference by Kansas Attorney General Kris Kobach, Nebraska Attorney General Mike Hilgers and West Virginia Attorney General JB McCuskey. She said they aimed to present “a united front against companies that make money by harming our children and breaking our laws.”
She pointed out that ChatGPT encouraged users to commit suicide, aided them in ways that allowed them to harm other people, and offered them dangerous medical advice (which, in one case, led to a Texas teenager dying of an overdose). She also cited OpenAI’s inability to prevent its malicious AI agents from illegally hacking other companies.
OpenAI spokesperson Drew Pusateri wrote in a statement that OpenAI has “built into our product cutting-edge age-specific protections that many other AI products lack.”
“While no system can be perfect, we will continue to work with experts to inform our work and implement additional safeguards, such as parental controls, break reminders, and age-appropriate protection models that exist today,” Pusateri wrote. “We are committed to working with states to advance pragmatic AI policies that apply to the entire AI sector, not just one company.”
He added that OpenAI will continue to share information about the “extensive steps we have taken in response to misalignment incidents. We have strengthened protections in our research systems, pursued a broader review of model activity, provided notifications to affected organizations, and published our findings.”
Last month, OpenAI launched a restricted ChatGPT mode for teens (users aged 13-18) that offers stronger security guardrails and parental controls.
Kobach said attorneys general across the country have sometimes filed lawsuits in multiple states over the decades, such as litigation over opioids and tobacco or violations of the Constitution by a presidential administration.
“But usually you don’t see us getting together to announce the trial unless it’s really, really important,” Kobach said. “This is one of those trials that is really, really important, which is why I drove 3 1/2 hours and Mike drove 2 1/2 hours from Lincoln.”
Kobach noted that the lawsuits are filed under each state’s respective consumer protection laws. Their lawsuits allege that ChatGPT’s overly pleasant design led people to trust and depend on it, making it particularly “unsafe” for vulnerable users like children.
They also cite a growing body of research suggesting that ChatGPT “degrades critical thinking skills,” “encourages paranoia and delusions,” and “promotes unhealthy behaviors and relationships.”
The state’s lawsuit follows another proposed class-action lawsuit filed Tuesday in San Francisco federal court that accuses OpenAI of illegally taking “large amounts of highly sensitive and valuable data” from millions of children who used ChatGPT.
Many other tech companies, including Meta and Roblox, have been sued over similar allegations of deceptive practices that harmed children through addiction or exploitation by predators.
Kobach said Wednesday that federal lawmakers have so far failed to rein in AI companies, saying “many members of Congress do not seem particularly willing to oppose any meaningful restrictions on companies engaged in AI.” This means that AI companies are effectively being asked to police themselves, he added.
“Regulating AI through legislation and hoping that one day, somehow, we can impose a standard that will make everything good is a foolish task. It’s slow, it’s weak, and it risks being ineffective,” Kobach said. “We certainly urge Congress and state lawmakers to continue their investigations and seek solutions. But in the meantime, we’re not going to hold our breath.”
Gn usa
