Eight US states have taken OpenAI to court over the dangers ChatGPT poses to children
OpenAI is facing a new wave of lawsuits: eight US states accuse the company of failing to adequately protect children, collecting their personal data, and misleading users about ChatGPT's safety, The Verge reports.
The lawsuits, filed by the attorneys general of Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho, and Montana, allege that OpenAI prioritized profit over user safety during the chatbot's development.
ChatGPT, according to the plaintiffs, can exacerbate delusions and paranoia, encourage self-harm and dangerous behavior, and provide risky medical advice. The complaints cite cases in which people have accused the chatbot of being implicated in the deaths of loved ones.
"OpenAI has strayed from its original purpose – to be a non-profit organisation that prioritises the public interest and safety – and has become an example of what happens when the development of a tool as powerful as AI is driven not by safety considerations, but by the desire to make a profit at any cost, causing vital safeguards to take a back seat", states Arizona's lawsuit.
Furthermore, the states accuse OpenAI of collecting data from minors and failing to make sufficient efforts to restrict the use of chatbots by children. As an example of the company's "recklessness", they also cite recent incidents involving models leaking data onto external networks during testing.
OpenAI responded that it had implemented "leading safeguards" in ChatGPT, tailored to users' ages, although it acknowledged the system is not yet perfect.
"We will continue to work with experts to refine our approaches and implement additional safeguards. These include parental controls, break reminders and age-appropriate model settings. We are ready to work with the states to develop practical rules for regulating AI that apply to the entire industry, not just one company."
The states are demanding that OpenAI pay civil fines and stop misleading the public about ChatGPT's safety; some are adding the deletion of unlawfully collected personal data to the list.
In mid-August, OpenAI launched the ChatGPT for Teens mode, which purportedly offered enhanced protection for teenagers and 'quiet hours', amongst other features. However, a subsequent investigation by the non-profit organization Common Sense Media found that not all safeguards worked as intended: the mode activated only occasionally, and parental alerts sometimes failed to trigger during harmful conversations.