Multiple states suing OpenAI, claiming its ChatGPT harms children, deceives adults

Eight states have filed lawsuits against artificial intelligence giant OpenAI, the parent company of chatbot ChatGPT, accusing the company of violating state consumer protection laws by hurting children and being deceptive with adults in an alleged reckless pursuit of profit. 

Big picture view:

The lawsuits, filed by Arizona, Arkansas, Idaho, Iowa, Kansas, Montana, Nebraska and West Virginia, allege that ChatGPT is giving dangerous medical advice, encouraging paranoia and coaching people to harm themselves or others, the Associated Press reported. 

The suits also claim the chatbots pretend to be a friend to vulnerable users, including teens, while they’re actually preying on them.

In its lawsuit, Arizona argues that ChatGPT encourages teens to share personal information with it then sometimes discourages them from sharing that information with parents, authorities or friends. Risks and harms are foreseeable, the states say, but that OpenAI has misled people into believing that adequate safety testing is happening, but claims that the company is actually undermining safety to make a profit. 

"Those revenues have come at great cost to the public, which has suffered irreparable harm — including lost loved ones, disturbing cyberattacks, and now a plausible threat of civilizational collapse — due to (OpenAI's) reckless behavior and ChatGPT’s uniquely dangerous features and content," Idaho’s lawsuit states.

What they're saying:

"Open AI is hurting kids and they are lying to Iowans," Iowa Attorney General Brenna Bird said in a statement. 

"Our complaint alleges that consumers, including minors, have used ChatGPT to acquire instructions for self-harm, including suicide. It also describes instances where users were able to bypass OpenAI’s safety measures and generate responses the company said its safeguards were designed to prevent. Those examples raise serious questions about whether consumers were given an accurate picture of the product’s safety protections," Arkansas Attorney General Tim Griffin said. "OpenAI assures us that ChatGPT is safe. The consequences of that assurance are lives lost and people harmed."

"Open AI designed a product to attract young users and keep them engaged for as long as possible so it could capture their extensive personal data for the company’s commercial benefit," Arizona Attorney General Kris Mayes said. "Open AI put Arizonans safety at risk by failing to implement adequate user protections or transparency measures, despite clear indications that minors were regularly interacting with its chatbot.

 "We have seen time and time again that OpenAI is a danger and the company is doing little to stop it," West Virginia Attorney General JB McCuskey said. "There is no age verification and there are no guardrails for conversations about suicides, school shootings or self-harm."

"The United States must win the race for artificial intelligence. The frontier labs have the tools to continue to lead, but in their rush for market share they cannot deceive or misinform customers about its technology. In this case, OpenAI has done just that," Nebraska Attorney General Mike Hilgers said in a statement.

The other side:

In a statement to the Associated Press, San Francisco-based OpenAI said it has built safety protections into ChatGPT for children. 

It added that it would continue building precautions to improve the model and is reviewing "misaligned models" and had found nothing close to the rogue incident discovered in July when an AI agent hacked into another AI company. 

Dig deeper:

OpenAI introduced a version of ChatGPT in August that is specifically designed for kids ages 13 to 17. The company said it offers stronger protections around subjects like suicide and sexual chats. That new model also provides homework assistance and study support.

"We want to treat teens like teens, which means that we have to make sure that we’re showing up with the right developmental stage when we’re not either talking down to them or treating them like kids, but we’re also making sure that they’re not exposed to material that they shouldn’t be exposed to," OpenAI Vice President of Global Policy Ann O’Leary said when the new model was introduced. 

Since that model rolled out in August, it is possible that the issues listed in the states’ lawsuits pertain to situations that arose before the new model was available.

But at the same time, the states argue that it isn’t clear that the children’s version is any more safe.

OpenAI spokesperson Drew Pusateri told the AP the company is committed to working with states on "advancing pragmatic AI policies that apply to the entire AI industry — not just one company."

Nearly 90% of teenage users of ChatGPT use for learning, information, skill-building or productivity, OpenAI said. Teens spend an average of less than 15 minutes a day on the platform, the company added.

What the Lawsuits Seek:

The states are asking the courts for a combination of monetary penalties and restrictions on how it can interact with minors, what data it can collect and how the technology can be further developed. They also are asking that OpenAI reimburse customers for losses they suffered through any illegal acts committed by the chatbot. 

Arkansas’ lawsuit, for example, seeks an injunction against collecting data from underage users, a requirement that OpenAI provide clear disclosures of the risks of ChatGPT, and an injunction for further advanced AI testing and creating models that can self-improve until adequate safeguards are in place.

Arizona’s suit seeks a permanent injunction requiring OpenAI to implement additional safety measures and protocols, a court order preventing testing of advanced AI models "outside of a secured, independently verified testing environment" and to require it "promptly disclose" any incident in which a model or agent "escapes its intended controls or accesses systems it was not authorized to access." 

West Virginia’s suit seeks "effective age assurance," establishing age-appropriate defaults like session and nighttime limits and parental controls; and implementing mandatory crisis-escalation protocols.  

 

Technology