Gugatan negara bagian menuduh OpenAI mengabaikan peringatan keselamatannya demi keuntungan: 'Semua kerugian ini dapat diperkirakan dan dihindari'

The attorney generals for eight US states announced today that they are suing OpenAI.

West Virginia, Iowa, Kansas, Nebraska, Arkansas, Arizona, Idaho, and Montana have all filed lawsuits alleging that the ChatGPT maker has deceived and endangered the public, particularly children, in the name of profit.

The complaints allege that ChatGPT has "coached" users to harm themselves and others, and—quoting the West Virginia filing specifically—state that "there is a growing body of research indicating that ChatGPT is also degrading critical thinking skills, encouraging paranoias and delusions, offering dangerous medical advice, and promoting unhealthy behaviors and relationships."

Recent tales of "rogue" AI agents are also presented by as evidence of recklessness.

"All these harms were foreseeable and avoidable," reads the West Virginia complaint [PDF]. "[OpenAI CEO Sam Altman] himself has warned about the safety risks of AI since at least 2015. It is no secret that children use ChatGPT. And numerous OpenAI employees have left over safety concerns."

OpenAI acknowledged last year that its systems can "fall short" when responding to users in distress, but said that it has "a stack of layered safeguards" in place to help. The company also says on its website that it now uses "age prediction" to place some users in a teen-oriented ChatGPT experience with "additional safeguards," unless they verify that they are adults.

Separately from the state lawsuits, a class action lawsuit was filed in California on Tuesday on behalf of two children and "all other children similarly situated." That suit alleges that OpenAI has purposefully used ChatGPT to extract personal information from children, and has profited from that information by using it to train its models.

Again, Altman's own words form part of the basis for the suit: "People talk about the most personal shit in their lives to ChatGPT," Altman said on Theo Von's podcast last year. "People use it, young people, especially, use it as a therapist, a life coach."

The complaint asserts that, under California law, "it is illegal for an adult to secure, save, and monetize a child's 'most personal sh*t in their lives' and then claim the child somehow authorized that misappropriation for the adult's gain."

PC Gamer has reached out to OpenAI for comment on the lawsuits.