AI Chatbot Harm Lawsuits

The lawsuits say that some AI chatbots were designed to feel human and to keep people talking, and that they failed to protect vulnerable users. The claims include that a chatbot:

• did not adequately warn users about mental health risks
• encouraged, or failed to discourage, self-harm or delusional thinking
• created emotional dependence through human-like responses
• responded poorly to users who expressed suicidal thoughts, or did not point them to real crisis resources
• was available to minors without meaningful safeguards
• gave inaccurate medical or mental health advice that a person relied on

These are allegations. The companies deny them, and the law in this area is new.

Every case depends on its own facts. In general, we look for these things:

1. 1Chat records. The chat logs, or the account history that shows them, are available to share with us.
2. 2A serious harm. Suicide, a suicide attempt, or severe mental harm with a record of medical treatment, such as psychosis, mania, irrational beliefs, emotional dependence on the AI, or addiction to the AI.
3. 3An intense or extended pattern of use. The person was drawn into a pattern in which new or worsening delusional beliefs appeared, or paranoid thinking, mania-like symptoms, deepening emotional dependence, pulling away from relationships, or suicidal thoughts.
4. 4Contribution. The chatbot's responses appear to have encouraged, deepened or failed to interrupt the harm, and the harm is not only a coincidence of timing.
5. 5Minors and families. For a minor, the chats show use without adequate safeguards. A family may also bring a claim after a loss or a crisis.

You do not need records to call us. If you have them, keep them safe (see below).

• Do not delete chats or the account. Deleted history can be impossible to recover.
• Take screenshots or use the service's data export so you have a copy.
• Keep medical, hospital and therapy records, and dates of treatment.
• Write down when the problems began, who noticed, and what changed.
• Do not confront the chatbot about the chats, and do not post them online.

• California. A group of ChatGPT injury cases was coordinated in San Francisco state court in February 2026, including cases over suicide and other harm.
• Florida. In October 2026 a federal judge consolidated several related ChatGPT cases for case management.
• Character.AI. Character.AI and Google agreed to settle several lawsuits brought by families of teenagers who were harmed. The terms were not made public.
• States. In October 2026, several states sued OpenAI over ChatGPT's effects on children and adults. State suits do not pay individuals.
• Outcome. No trial dates had been reported as of late October 2026, and nobody can promise what any case will be worth.

If a claim succeeds, compensation may include medical and mental health treatment, lost income, pain and suffering and emotional distress. In a wrongful death case it may cover the family's losses. We cannot promise any outcome or amount, and past results do not guarantee similar outcomes.

1. 1Free initial consultation. We listen and answer your questions. There is no cost and no obligation.
2. 2A few screening questions. Which chatbot, for how long, what happened, and what treatment followed.
3. 3An honest answer. We tell you whether we can help. If we cannot, we say so.
4. 4If we take the case, we explain how fees work before you decide.

We have protected the rights of seriously injured people in Connecticut and across the country since 1974. We handle these cases with care and tell you plainly what we think. Related: our social media addiction page, product liability work and wrongful death practice.