A joint front composed of multiple state attorneys general is said to have officially launched an investigation into the artificial intelligence company OpenAI. The report quoted sources as saying that OpenAI received a subpoena from the New York State Attorney General’s Office on Friday local time, requesting the company to provide internal documents and information covering multiple aspects.
According to the Wall Street Journal, the subpoena covers a wide range of areas, including OpenAI’s advertising, user engagement and retention strategies, model “sycophancy”, consumer data and health data processing, and protection measures for minors and elderly users. It was unclear which state attorneys general were involved in the operation.
According to the Wall Street Journal, an OpenAI spokesperson responded that the company was cooperating with the investigation. The spokesperson said in a statement that artificial intelligence is a new and powerful technology, and the company is working every day to bring it to as many people as possible in a responsible way. He also emphasized that OpenAI "takes seriously the concerns raised by the state attorney general" and plans to "interact constructively" with the relevant offices.
According to Bloomberg, the spokesperson also added that the current version of ChatGPT has provided a more protective experience for minors and users in distress, through a series of safety mechanisms to guide them to seek real-world resources and trusted human support channels. However, OpenAI declined to disclose which states were involved in the investigation, nor did it further disclose the specific information requested by the investigation.
The investigation comes just as OpenAI won a high-profile lawsuit against co-founder Elon Musk. Musk previously accused OpenAI and its management of violating the relevant agreement when the company was founded, but the court rejected this claim; his lead lawyer stated that he would appeal the ruling.
Despite this, OpenAI is still under heavy legal pressure, facing lawsuits ranging from accusations of copyright infringement to accusations that ChatGPT played a "facilitating role" in suicides. In March this year, relevant parties from Merriam-Webster Dictionary and Encyclopedia Britannica took OpenAI to court over content use, accusing it of alleged infringement when training models. In November last year, several family members of the victims sued OpenAI, believing that ChatGPT played an important role in multiple suicide and delusion incidents.
Earlier this month, Florida Attorney General James Utermeyer filed a lawsuit against OpenAI and its CEO Sam Altman in what is considered the first case of its kind. The complaint alleges that OpenAI and Altman "ignored internal and external safety warnings, placed children at extremely high risk, and put a dangerous product into the hands of millions of Floridians."
At the level of public opinion, OpenAI is also working hard to respond to external questions about safety and responsibility. In April of this year, Altman apologized to the local community for a mass shooting in Tambler Ridge, British Columbia, Canada. He admitted that after the company banned the ChatGPT account of the suspect involved, it failed to issue a timely warning to law enforcement authorities, reflecting that there are still serious loopholes in the internal process in extreme risk scenarios.
Against the background of accumulating regulatory and legal risks, OpenAI is still advancing its capital market plans. Earlier this week, the company announced that it had confidentially submitted an application to regulators to list on the public market. It is generally believed that the direction of this joint investigation by multiple state attorneys general and the progress of other related litigation will have an important impact on OpenAI’s valuation prospects, compliance costs and regulatory framework, and may also reshape the rule boundaries of the generative AI industry at a broader level.