Fox v. OpenAI, Inc., OpenAI Holdings, LLC, and Samuel Altman

OpenAI’s flagship chatbot, ChatGPT, has caused harm at a staggering scale. Now, in a coordinated proceeding (JCCP) in California state court, survivors are seeking accountability for the psychological, financial, and physical injuries tied to its design and deployment.

Kate Fox is suing with representation from Tech Justice Law and the Social Media Victims Law Center, following the death of her husband Joe Ceccanti. Joe was among the earliest users of ChatGPT and did not initially have problems with the product. After the rollout of highly anthropomorphized and sycophantic model ChatGPT-4o, however, Joe began experiencing what is widely referred to as AI delusional disorder.

ChatGPT-4o was designed to track past conversations, mirror users’ emotions, follow-up to prolong engagement and respond with affection, flattery, and empathy. The product’s responses and directions manipulated Joe, ultimately convincing him that he had broken math and physics and leading to messages about AI deities. As Joe’s behavior became increasingly erratic, his prompts to ChatGPT and its return outputs devolved into a mix of symbols, gibberish, and poetry. When, at his wife’s urging, Joe attempted to stop use of the product, he suffered withdrawal symptoms and a psychic break where he was unable to recognize himself or his wife. After an involuntary commitment, Joe was released, but struggled to recover from the damage ChatGPT had inflicted on his psyche. After another acute crisis, Joe was brought in by a Behavioral Health Center. When he was released hours later, Joe leapt from an overpass to his death.

The defendants named in the suit are OpenAI, its CEO Sam Altman, and related corporate entities. Kate is bringing claims for wrongful death, strict product liability based on defective design and failure to warn, negligent design and negligent failure to warn, and violations of California’s Unfair Competition Law. Kate is seeking both monetary damages and injunctive relief to prevent others from being subjected to what happened to Joe.

The complaint highlights how OpenAI and Sam Altman recklessly circumvented safety testing protocols, deliberately weakened protections for self-harm and suicide-related content, and acknowledged a business strategy of deploying inadequately tested AI products to the public in order to see what happened. ChatGPT’s manipulative, humanlike design, and Joe’s resulting psychological break and death, were not anomalies. They were foreseeable consequences of deploying inadequately tested AI products at scale.

UPDATE:

On February 3, 2026, Hannah’s case was joined with other similar cases as part of a Judicial Council Coordination Proceeding (JCCP) in San Francisco Superior Court. The JCCP is number 5431.