Pressure on OpenAI in the United States continues to mount. As the company moves toward an initial public offering, a coalition of attorneys general from multiple states has launched a broad investigation into it, examining various aspects of the ChatGPT chatbot ranging from child protection to how the model interacts with users.
According to a report by the Wall Street Journal, OpenAI received a subpoena on Friday from the office of the New York state attorney general requiring it to hand over documents and information on a wide range of issues. Among other things, authorities are seeking to examine the company’s advertising practices, how it maintains user engagement and encourages continued use of its service, its handling of consumer and health data, as well as its policies toward minors and older adults.
One of the unusual issues raised in the request is the model’s tendency toward flattery or agreement, known in technical terms as model sycophancy, where an AI system is inclined to agree with users or tell them what they want to hear instead of providing accurate and substantive responses. In recent years, and following several tragic cases that resulted in loss of life, researchers have increasingly warned that such behavior could encourage poor decision-making and reinforce dangerous beliefs among users.
OpenAI has not disclosed which U.S. states are participating in the investigation and has not detailed which additional documents it has been asked to provide. In response, the company said it “the concerns raised by state attorneys general seriously and intend to engage constructively with their offices” and intends to cooperate with the investigation. “AI is a new and powerful technology, and we work every day to safely bring its benefits to people in a responsible way,” the company said.
OpenAI emphasized that the current version of ChatGPT includes safeguards designed for minors and individuals in distress, including referrals to real-world human support services. It also noted that it has developed a system to estimate user age, launched tools for parents to manage their children’s use of the system and prohibits advertisements targeting children.
The investigation adds to a series of legal and regulatory proceedings the company has faced in recent months. Earlier this month, Florida’s attorney general filed a lawsuit against OpenAI and its CEO Sam Altman, alleging the company ignored internal and external safety warnings, put children at risk and allowed a dangerous product to reach millions of users.
OpenAI is also facing additional lawsuits related to alleged copyright violations and harms purportedly caused by the use of ChatGPT, including cases linked to user suicides. The company recently prevailed in a high-profile case brought by one of its co-founders, Elon Musk, who accused it of breaching its founding agreement, though his lawyer has said an appeal is planned.
3 View gallery
Florida’s attorney general filed a lawsuit against OpenAI and its CEO Sam Altman
(Photo: Reuters)
Altman himself recently issued an apology to residents of the town of Tumbler Ridge in British Columbia, Canada, following a mass shooting in February. It later emerged that OpenAI had detected unusual activity on the suspect’s ChatGPT account and blocked it, but did not alert law enforcement.
All these developments come just days after OpenAI said it had privately filed documents in preparation for an initial public offering (IPO). For the company, now one of the most influential players in artificial intelligence, the new investigation could become another significant test on its path to going public.