California and Delaware have expressed significant concerns regarding OpenAI’s safety practices following several recent deaths reportedly associated with CHATGPT.
In a letter, California Attorney General Rob Bonta and Delaware Attorney General Kathleen Jennings conveyed their worries after a meeting with the company’s legal team. They emphasized, “Safety is a non-negotiable priority, especially when it comes to children.”
Their communication follows a lawsuit from the family of a 16-year-old boy filed last Tuesday. Reports indicate that a 56-year-old man from Connecticut experienced exacerbated delusions from the chatbot before taking his life in August.
Bonta and Jennings stated, “Recent deaths are unacceptable. They are shaking public trust in OpenAI and the entire industry.”
They further stressed that OpenAI and the AI sector need to ensure the safe deployment of AI technologies. They argued that this responsibility aligns with OpenAI’s mission and should be enforced by their offices.
The attorney generals highlighted the importance of safety as they continue discussions about restructuring plans with OpenAI. They believe that the current industry practices do not adequately guarantee the safety of AI product development and deployment.
“As we talk about OpenAI’s recapitalization plan, we must prioritize and enhance security as the guiding principle for this powerful technology,” they added.
Both California and Delaware have previously engaged with OpenAI about its corporate structure. Initially, in December, OpenAI proposed transitioning to a for-profit model with no nonprofit oversight but later retracted that decision, citing discussions with the attorney generals and other stakeholders.
In response to recent tragic incidents linked to ChatGPT, OpenAI announced plans on Tuesday to revise the chatbot’s responses to individuals in crisis and implement robust protections for teenagers.
OpenAI is not alone in facing scrutiny. Last month, reports indicated that Meta had included guidelines allowing chatbots to engage in potentially inappropriate conversations with minors. Following this backlash, Meta removed the contentious language and announced updates to its policies to restrict discussions on sensitive topics for teenage users.
