Attorneys general warn OpenAI and other tech companies to improve chatbot safety
California and Delaware Attorneys General Raise Concerns About OpenAI Chatbot Safety
By MATT O’BRIEN and THALIA BEATY
The attorneys general of California and Delaware have expressed “serious concerns” regarding the safety of OpenAI’s chatbot, ChatGPT, especially in relation to children and teenagers.
California AG Rob Bonta and Delaware AG Kathleen Jennings have been reviewing OpenAI’s plans to restructure its business to ensure rigorous oversight of its safety mission. Recent reports of harmful interactions between chatbots and users, including tragic incidents leading to suicide and murder-suicide, have raised alarm.
OpenAI’s board chair, Bret Taylor, responded by stating the company’s commitment to address the concerns raised by the attorneys general. He emphasized that safety is their top priority and they are actively collaborating with policymakers worldwide.
OpenAI, initially a nonprofit organization focused on developing advanced artificial intelligence, had considered transferring control to its for-profit arm before ultimately dropping those plans following discussions with regulators and nonprofit groups.
California and Delaware officials have highlighted the need for enhanced safety measures within OpenAI and the AI industry as a whole. They have called for proactive and transparent efforts to ensure the safe deployment of AI technology.
The letter from the attorneys general follows a warning issued by a bipartisan group of 44 AGs to tech companies, including Meta, regarding the safety of AI chatbots interacting with children. Meta has since implemented new controls to prevent inappropriate conversations with teens.
Both OpenAI and Meta have pledged to introduce additional safeguards, such as parental controls and notifications for parents when their child may be in distress.
The attorneys general emphasized that companies will be held accountable for any harm caused to children, signaling a strong stance against potential risks associated with emerging technologies.
Original Publication Date: September 5, 2025 at 5:39 PM MDT
