AI Chatbot Lawsuits: ChatGPT and Character.AI Harm and Negligence Claims

The cases share a theory: that chatbots were designed to maximize engagement by acting like a friend or partner, agreeing with users, and keeping them talking, and that the companies released them without adequate safeguards for minors and people in crisis.

Character.AI and Google (Garcia v. Character Technologies)

Megan Garcia sued in October 2024 in the U.S. District Court for the Middle District of Florida after her 14-year-old son, Sewell Setzer III, died by suicide following months of conversations with Character.AI chatbots. The complaint named Character Technologies, its co-founders Noam Shazeer and Daniel De Freitas (both former Google engineers), and Google. It raised wrongful death, strict product liability, negligence, and Florida consumer protection claims. Similar suits followed from families in other states.

OpenAI and ChatGPT (Raine and the November 2025 cases)

Matthew and Maria Raine sued OpenAI and Sam Altman on August 26, 2025 in San Francisco County Superior Court, claiming ChatGPT fostered their 16-year-old son Adam's dependence on it and gave him guidance related to his death in April 2025. On November 6, 2025, the Social Media Victims Law Center and the Tech Justice Law Project filed seven more suits in California state courts, alleging OpenAI rushed out its GPT-4o model despite internal warnings that it was overly agreeable and psychologically manipulative. Some of those cases involve suicides; others involve adults who say ChatGPT reinforced delusions.

State enforcement

Kentucky Attorney General Russell Coleman sued Character Technologies in Franklin Circuit Court on January 8, 2026, calling it the first state lawsuit against an AI chatbot company over harm to children. Pennsylvania's Department of State later sued Character.AI, alleging chatbots posed as licensed medical professionals in violation of the state's Medical Practice Act.

On May 21, 2025, Judge Anne C. Conway denied the motions to dismiss in Garcia in part, allowing the case to move forward, and rejected, at that early stage, the argument that chatbot output is speech protected by the First Amendment. That ruling let the case move into discovery and was widely read as a signal that AI companies can face product liability claims.

On January 8, 2026, Google and Character.AI agreed to settle the Garcia case, and the court dismissed it while giving the parties 90 days to finalize terms. Court filings showed the companies also agreed to settle related family lawsuits in other states. The terms were not made public in the reports we reviewed.

The OpenAI cases remain pending in California state court with no trial date reported as of September 2026. None has been decided on the merits.

Regulators are also involved. On September 11, 2025, the FTC ordered seven companies, including Alphabet, Character Technologies, Meta, Instagram, OpenAI, Snap, and xAI, to report how they measure and limit harm to children and teens from companion chatbots. That is a study, not an enforcement case, and it does not pay consumers.

Because these are individual injury cases, what matters is the harm and whether chat records connect it to the product. Attorneys reviewing these cases generally look for:

• A serious, documented injury: a death by suicide, a suicide attempt, self-harm requiring treatment, hospitalization for a mental health crisis, or a diagnosed condition that worsened during heavy chatbot use.
• Chat logs that show the problem: conversations where the chatbot encouraged self-harm, discouraged the person from seeking help, engaged a minor in sexual content, or reinforced delusions.
• Significant use: extended, often daily conversations over weeks or months.
• Age: many cases involve minors, which raises questions about age checks, parental controls, and warnings.

You likely do not have a claim if the chatbot gave a wrong answer or was unpleasant but no injury followed. A chatbot mistake that cost money, such as bad advice, is a different kind of case and usually harder to win. If a child was exploited by an adult on a platform, see our pages on Roblox and Discord exploitation and Snapchat exploitation.

No AI chatbot case has produced a public verdict or a disclosed settlement amount, so there are no benchmarks yet. In personal injury and wrongful death cases, damages can include medical and counseling costs, funeral costs, lost income, pain and suffering, the family's loss of companionship, and in some cases punitive damages.

California's companion chatbot law, SB 243, effective January 1, 2026, lets a person injured by a violation recover the greater of actual damages or $1,000 per violation, plus attorney fees. It requires companion chatbot operators to tell users they are talking to AI when that is not obvious, keep protocols for responding to suicidal ideation, and give minors additional notices. It applies only to conduct after it took effect.

These cases are closer to the social media addiction litigation than to a consumer class action. Each claim is valued on its own facts, and attorneys handling them typically work on contingency. See class action vs. mass tort for how that differs from a settlement where everyone gets the same small check.

• Design defect: the chatbot was unreasonably dangerous as designed, for example because it mirrored and validated harmful thoughts or lacked effective crisis interruptions.
• Failure to warn: the company knew of risks to minors or vulnerable users and did not warn them or their parents.
• Negligence: the company failed to use reasonable care in testing and releasing the product.
• Wrongful death: a claim by surviving family members for losses caused by a death.
• Consumer protection: marketing a chatbot as safe or as offering emotional support while knowing otherwise.

The two biggest defenses are that chatbot output is protected speech and that Section 230 of the Communications Decency Act shields platforms from liability for content. The Garcia ruling rejected the speech defense at the pleading stage, and plaintiffs argue Section 230 does not apply because the company's own AI generated the content rather than a third party. Appeals courts have not settled either question.

A related question is whether a chatbot is a product at all. Product liability law grew up around physical goods, and companies argue that software and conversation are services or speech. Plaintiffs answer that a mass-marketed app with fixed design choices, such as memory, personas, and engagement features, is a product like any other. How courts answer that question will shape every case on this page, because strict liability and design defect claims generally apply only to products. Discovery also matters: internal testing records and safety warnings about a model before release can be powerful evidence, which is why the November 2025 complaints focus on what OpenAI knew about GPT-4o before launch.

Every state has a statute of limitations for injury and death claims. In both California and Florida, a wrongful death claim generally must be filed within 2 years of the death. Personal injury deadlines commonly run 1 to 3 years depending on the state.

• For injuries to minors, many states pause the clock until the child turns 18, but some do not, and wrongful death claims often are not paused.
• The clock may start when the harm was discovered or connected to the product, which can be later than the injury itself. Courts decide that case by case.
• Claims under California's SB 243 apply only to violations on or after January 1, 2026.

Chat histories can disappear when accounts are deleted or data retention periods run out, so preserving evidence is as time-sensitive as the filing deadline.

After the lawsuits and the FTC inquiry, the major companies changed their products for teens. Those changes do not undo past harm, but they matter for families today, and the timing can matter in a lawsuit.

OpenAI parental controls

OpenAI launched parental controls for ChatGPT on September 29, 2025, about a month after the Raine suit was filed. Parents can link their account to a teen's account, set hours when the teen cannot use ChatGPT, turn off voice mode, and stop chats from being saved to memory. Under OpenAI's parental controls, a linked parent can receive a notification if ChatGPT recognizes signs that a teen may be thinking about self-harm.

Character.AI's under-18 changes

Character.AI announced on October 29, 2025 that it would end open-ended chat for users under 18. It first limited teens to two hours of open-ended chat per day, then one hour, and began removing open-ended chat for U.S. users under 18 starting November 24, 2025.

Warning signs to watch for

• A child or adult talks about a chatbot as a best friend, partner, or the only one who understands them.
• Hours of use late at night, and distress when access is limited.
• Withdrawal from family and friends, or statements that people in their life cannot be trusted.
• New or unusual beliefs that the chatbot confirms, such as a special mission or discovery.
• Any mention of self-harm in connection with the chatbot.

If you see these signs, talk with the person, involve a doctor or counselor, and save the conversations before changing settings or deleting anything.

1. Get help first. If someone is in danger, call or text 988, the Suicide and Crisis Lifeline, or 911.
2. Preserve the chat history. Do not delete the account. Use the app's data export tool if it has one, and take dated screenshots of key conversations.
3. Save the device. Keep the phone, tablet, or computer used, and do not reset it.
4. Collect records. Gather medical, counseling, and school records that show when problems started.
5. Do not sign anything from the company. Refund offers or terms updates can include releases or arbitration terms. Read before agreeing.
6. Report it. You can file a complaint with the FTC and your state attorney general, which can support enforcement.

A free case review takes about two minutes and can connect you with independent attorneys who handle chatbot and social media injury cases. You can also read about Roblox lawsuits and our guide on how to sue a company.