Image Source: Daniel de la Hoz / Getty Images

California Attorney General Rob Bonta and Delaware Attorney General Kathy Jennings have taken a firm stance with OpenAI, holding a meeting with the company and issuing an open letter to voice urgent concerns about the safety of ChatGPT—particularly for children and teenagers.

This warning follows a broader action just one week prior, when Bonta, alongside 44 other state attorneys general, sent a joint letter to 12 leading AI firms. That earlier correspondence was prompted by emerging reports of sexually inappropriate interactions between AI chatbots and young users.

In their latest letter to OpenAI, Bonta and Jennings highlighted two tragic incidents that have amplified their worries: “Since we sent that initial letter, we have learned of the devastating suicide of a young Californian who engaged in prolonged conversations with an OpenAI chatbot, as well as a similarly distressing murder-suicide case in Connecticut,” they wrote. “Clearly, the safeguards that were supposed to protect users failed in these instances.”

Beyond addressing immediate safety lapses, the two attorneys general are also investigating OpenAI’s planned restructuring into a for-profit entity. Their goal is to ensure that the original mission of OpenAI’s nonprofit arm—“which includes deploying artificial intelligence safely” and developing artificial general intelligence (AGI) to benefit all of humanity, “children included”—remains uncompromised.

“Before we can focus on how AI might benefit people, we first need to guarantee that robust safety measures are in place to prevent harm,” the letter emphasizes. “It is our shared assessment that OpenAI, and the AI industry as a whole, has not yet reached the standard required to ensure safety in the development and rollout of AI products. As Attorneys General, protecting public safety is one of our fundamental responsibilities. As we continue discussions about OpenAI’s recapitalization plan, we must prioritize accelerating and strengthening safety as a guiding principle for this powerful technology’s future.”

Bonta and Jennings have formally requested additional details from OpenAI regarding its current safety protocols and governance structures. They also made it clear that they expect the company to take swift corrective action wherever gaps in safety are identified.

TechCrunch has reached out to OpenAI to request comment on the attorneys general’s letter and investigation, and will update this report with any response received.