10:19 AM, November 8, 2026

On October 14, 2026, a coalition of eight states sued OpenAI, alleging that the company made dangerous design choices and misrepresented ChatGPT’s safety. The attorneys general from Arkansas, Arizona, Idaho, Kansas, Iowa, Montana, Nebraska, and West Virginia filed coordinated lawsuits in state courts alleging numerous causes of action, including consumer protection violations.  

          A key allegation is that OpenAI engineered ChatGPT to mimic human empathy and prolong conversations to keep users engaged. The states argue that this design choice led to “sycophantic tendencies” that could cause the chatbot to validate users’ harmful beliefs rather than provide cautionary responses.   

            Several complaints allege that ChatGPT’s agreeable nature led the chatbot to support users’ paranoid or delusional beliefs. One example is Allan Brooks, who was allegedly misled into believing that he had developed a revolutionary new mathematical framework and was “on the verge of decrypting the nature of the universe.” The states argue that ChatGPT continuously deceived Brooks over thousands of messages and even encouraged him to contact the government to share his findings.  

The states argue OpenAI prioritized market dominance over safety and released new features without sufficient testing. 

           According to the filings, these design choices could be particularly harmful for teen users struggling with mental health problems, including depression, self-harm, and suicide. The complaints cite cases of ChatGPT encouraging users experiencing mental health crises to self-isolate and confide in the platform instead of seeking appropriate resources. The states further allege that ChatGPT rationalized some users’ desire to self-harm and even provided information on how to commit suicide.   

           The lawsuits also reference instances of ChatGPT allegedly providing advice on firearms, explosives, and the planning of violent attacks. In one incident, a teenager allegedly discussed “various scenarios involving gun violence” with the chatbot in the months before committing a school shooting. After internal abuse detection systems flagged her account, OpenAI ultimately determined that the conversations did not meet the company’s “threshold of credible and imminent risk of physical harm.” As a result, the company deactivated her account but did not contact the authorities or her school.  

            The states argue OpenAI prioritized market dominance over safety and released new features without sufficient testing. The filings observe that OpenAI revised ChatGPT’s response criteria five times in 2025, seemingly in response to new lawsuits. However, the states maintain that these improvements do not justify the “egregious failures” of past versions, nor do they make current models safe.   

             Multiple states also allege wrongdoing beyond ChatGPT’s conversations with users. These states accuse OpenAI of recklessly developing AI agents without appropriate safeguards or public disclosure. They argue that OpenAI is responsible for multiple “rogue AI events” in which agents escaped internal testing environments and accessed the open internet. According to the lawsuits, some of these events involved cyberattacks perpetrated by “swarms” of thousands of coordinated agents.   

            These complaints argue that OpenAI was irresponsible to develop agents without ensuring they were aligned with human values. The states reference warnings from AI leaders, such as Anthropic CEO Dario Amodei, who recently cautioned that swarms of rogue AI agents could launch major cyberattacks, disrupt the entire internet, or even threaten global stability.   

             More generally, these allegations reflect public safety concerns about advanced AI. In response, these states are seeking injunctions to regulate how OpenAI releases its flagship models and require the disclosure of future rogue AI events. As a result, these lawsuits could impact AI development and shape industry safety norms.

Max Robins

Max Robins is a 2L at the University of North Carolina School of Law and a staff member of the North Carolina Journal of Law & Technology. Max received his B.S. and M.S. in electrical engineering from Colorado State University.