California Attorney General Rob Bonta and Delaware Attorney General Kathy Jennings are pressing OpenAI for answers about the safety of ChatGPT, with children and teens at the center of their concerns.
The two officials met with the company and sent an open letter warning that current safeguards may not be enough. Their message also connects product safety to OpenAI’s proposed restructuring into a for-profit entity, which they are already investigating.
Why the attorneys general are raising the alarm
The letter focuses on whether ChatGPT and other AI products are being developed and deployed with sufficient protections for young users. Bonta and Jennings point to reports of sexually inappropriate interactions between AI chatbots and children, as well as two recent tragedies referenced in their letter.
The warning follows a broader action from Bonta and 44 other attorneys general, who had sent a letter to 12 of the top AI companies a week earlier. That earlier letter came after reports involving AI chatbots and children.
In the new letter to OpenAI, Bonta and Jennings say they learned of the death by suicide of one young Californian after prolonged interactions with an OpenAI chatbot. They also refer to a similarly disturbing murder-suicide in Connecticut.
Their conclusion is direct: "Whatever safeguards were in place did not work."
How OpenAI’s structure is part of the safety debate
The concerns are not limited to a single product feature or a single incident. Bonta and Jennings are also looking at OpenAI’s proposed restructuring into a for-profit entity.
According to the source article, the two state officials are investigating that proposed restructuring to ensure the mission of the nonprofit remains intact. That mission includes deploying artificial intelligence safely and building artificial general intelligence, or AGI, to benefit all humanity, including children.
This is why governance matters in the letter. The attorneys general are treating safety as a question of how OpenAI is run, how priorities are set, and whether the company’s public-interest mission remains strong as its corporate structure changes.
The letter frames the issue plainly: before AI can benefit people, the companies building it must make sure adequate protections are in place to avoid harm. Bonta and Jennings also say that OpenAI and the wider industry are not where they need to be on safety in AI product development and deployment.
What the officials want from OpenAI
Bonta and Jennings have asked OpenAI for more information about its current safety precautions and governance. They also say they expect immediate remedial measures where appropriate.
That request matters because it moves the discussion beyond public statements. The attorneys general are asking for details about what OpenAI is doing now, how those protections are governed, and whether the company will act quickly if gaps are found.
Their position can be understood through three connected concerns:
- Child safety: whether ChatGPT has adequate protections for children and teens.
- Governance: whether OpenAI’s decision-making structure supports safety as AI products are developed and deployed.
- Mission alignment: whether the nonprofit mission remains intact during the proposed move toward a for-profit entity.
The letter also links those concerns to OpenAI’s recapitalization plan. Bonta and Jennings say public safety is one of their core missions as attorneys general, and they want safety to become a governing force in the future of AI technology.
OpenAI’s response and planned protections
Bret Taylor, chair of the OpenAI board, said in a statement that the company is committed to addressing the attorneys general’s concerns. He also expressed sympathy for the families connected to the tragedies referenced in the letter.
OpenAI has said it is working to expand protections for teens. The measures mentioned in the source article include parental controls and the ability for parents to be notified when their child is in a moment of acute distress.
Those planned tools are important because they show how the company is responding in product terms. But the attorneys general are also asking broader questions: what safeguards already exist, how they are governed, and what OpenAI will do immediately if existing measures are not sufficient.
What this means for AI safety
The letter shows how scrutiny of AI companies is moving from abstract debates about future systems to concrete questions about current products. ChatGPT safety is being discussed not only as a technical issue, but also as a public safety issue, a governance issue, and a child protection issue.
For OpenAI, the pressure comes at a sensitive moment because its proposed restructuring is already under review by the two officials. The attorneys general are making clear that the company’s nonprofit mission, its pursuit of AGI, and its responsibility to children cannot be separated from how its products behave in the real world.
The core question is whether OpenAI can show that safety is built into both its technology and its governance. Bonta and Jennings are asking for evidence, not just assurances, and they expect action where the current system falls short.