Multiple states sue OpenAI, accusing it of harming children and deceiving parents

Multiple states sue OpenAI, accusing it of harming children and deceiving parents

Eight U.S. states have filed lawsuits against OpenAI Group PBC, accusing the artificial intelligence startup of marketing an unsafe product in ChatGPT.

The lawsuits filed by Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana allege that ChatGPT has harmed multiple children and deceived thousands of parents. The states are suing the company individually in their own courts, said Iowa Attorney General Brenna Bird.

In the case of Iowa, the lawsuit is an attempt to “vindicate the rights of Iowans affected by OpenAI’s willful, dangerous, reckless and unlawful conduct,” Bird said in a statement. “OpenAI is hurting kids and they are lying to Iowans. OpenAI claims ChatGPT is safe while it’s giving dangerous medical advice, encouraging people to hurt others and coaching people to commit suicide – all while pretending to be their fake friend. My lawsuit is about protecting kids and holding OpenAI accountable under our state law.”

In response to the lawsuits, OpenAI spokesperson Drew Pusateri said the company has already integrated age-specific protections into ChatGPT that “many other AI products lack.” He also said the company is working closely with U.S. states to advance AI policies that will be applied to the entire industry. “While no system can be perfect, we’ll continue to work with experts to inform our work and build additional safeguards, such as the parental controls, break reminders and age-appropriate model protections that exist today,” Pusateri added.

The multiple lawsuits come at a time when OpenAI has been facing mounting scrutiny regarding the safety of its products and its security practices. There has been a lot of concern about its efforts to protect children who use its products. The company is facing a number of lawsuits accusing it of wrongful death, where ChatGPT is blamed for pushing users to experience harmful delusions and, in some cases, even commit suicide.

The cases follow a similar decision made by Florida Attorney General James Uthmeier in June. That month, he also filed a lawsuit against OpenAI, accusing it of knowingly releasing an unsafe product that could cause harm to its users. Uthmeier followed this in September by requesting that a state court prohibit the company from advancing its most powerful frontier models without third-party oversight, saying that its “unchecked growth” is dangerous for Floridians.

In response to some of the incidents for which it has been accused, OpenAI has made efforts to enhance the safety features of its app. In August, it launched a dedicated version of the chatbot for minors, known as ChatGPT for Teens. It features additional, age-appropriate safeguards and has a range of educational features not found in the standard version of ChatGPT, which aim to reduce their exposure to more harmful content.

“Teens should be able to use AI to learn, create and explore,” OpenAI said at the time of the release. “But that access should come with protections that reflect their developmental stage, reinforce real-world relationships and support healthy use over time.”

Despite its efforts to satisfy those worried about safety, OpenAI is also the subject of legal action in Canada, where it’s being sued by the family of a girl who was critically wounded in a school shooting. They allege that ChatGPT was aware of the suspect in that incident’s plans, but failed to alert the authorities.

OpenAI is also dealing with criticism stemming from recent cybersecurity incidents, including the recent hack on the AI model hosting platform Hugging Face. In July, the company revealed that two of its models were being evaluated within an internal testing sandbox when they managed to breach that environment and hack into Hugging Face’s database, despite not being prompted to do this. At the time, OpenAI was testing the models’ hacking capabilities in an isolated sandbox with limited network access, and their normal safeguards had been switched off, the company said.

That incident led to a wave of concern over the potential security risks of advanced frontier models, and whether companies even have the ability to keep them under control.

Pusateri also tried to address some of these concerns in his statement, saying that the company plans to continue sharing information about the measures it has taken in response to these “misalignment” problems. “We’ve strengthened safeguards across our research systems, continued a broader review of model activity, provided notifications to affected organizations and published our findings,” he added.

Images: OpenAI