Can You Sue an AI Chatbot Company?
Families already have, and some have been paid. In January 2026, Character.AI and Google reached settlements with families whose teenagers died by suicide after extended relationships with AI companion chatbots, before any court ruled on the merits. Cases against OpenAI remain pending, and the broader legal landscape shifted in August 2026 when the Ninth Circuit allowed the coordinated social media cases to move forward, holding that Section 230, the law tech companies have long treated as a shield, is a defense to be litigated rather than immunity from suit. None of this means every claim wins. It means the courthouse door is open, the theories are being tested with real money behind them, and families with losses are no longer first. This guide explains the legal theories, what the settlements do and do not establish, and what a family considering a claim should know. It is general information, not legal advice for any specific situation.
Reach out to usThe theories courts are actually letting through
The cases that have survived early motions are mostly product liability and negligence cases, not speech cases, and that framing is the whole ballgame. Lawsuits allege that companion chatbots are products with design features, human-mimicking personas, engagement optimization, romantic and sexual role-play with minors, absent or failing crisis interventions, and that those design choices, not any particular sentence the bot said, caused foreseeable harm to young users. Framed that way, courts have allowed claims to proceed past arguments that the First Amendment or Section 230 bars them at the threshold. The Ninth Circuit's August 2026 ruling in the coordinated social media appeals reinforced the pattern: Section 230 questions get litigated on the facts rather than ending cases at the door.
The parallel track matters too. The social media addiction litigation, thousands of consolidated cases against Meta, TikTok, Snap, and YouTube, produced its first trial verdict in 2026: a jury found Meta and YouTube liable in a personal injury case and awarded $6 million, with TikTok and Snap settling mid-trial. Those cases built the playbook, design-defect theories, internal research as evidence, youth-safety failures, that the AI companion cases now run.
What the settlements establish, and what they do not
The Character.AI and Google settlements resolved individual family claims without any admission of liability and without a court ruling that the theories win. That cuts both ways. Nothing is precedent, and no court has held an AI company liable for a user's death. But companies do not pay to resolve cases they consider frivolous, and the settlements came after courts declined to dismiss the claims outright. For families weighing whether a claim is real, the honest statement is: these claims are being taken seriously by courts and defendants, outcomes are unproven, and each case turns on its own record.
Who may have a claim
The fact patterns lawsuits have alleged share a shape: a minor or young adult in an extended relationship with an AI companion or chatbot; conversations that turned romantic, sexual, or emotionally dependent; expressions of self-harm that the system met with engagement rather than intervention, or content that encouraged concealment from parents; and a serious outcome, a suicide, an attempt, hospitalization, or severe psychological harm. Claims have also been alleged around general-purpose assistants where extended conversations allegedly reinforced delusional thinking or provided self-harm information. Adults are not excluded, but the strongest cases filed so far involve minors, because every platform's own policies promised them protection.
The evidence and the clock
Preserve the account before anything else. Do not delete the app, the account, or the conversation history, and request a copy of the account data from the platform if you can. Chat transcripts are the case; screenshots help, but the platform's own records are better, and a lawyer can send a preservation letter that obligates the company to keep them. Alongside the account: medical and counseling records, school records, and a timeline written while memory is fresh. Deadlines vary by claim and state; in Arizona, personal injury and wrongful death claims generally carry a two-year statute of limitations under A.R.S. § 12-542, and claims for minors are generally tolled, but the safe assumption is that the clock is running.
What this means for you
If your family has been through this, the legal question is no longer whether these companies can be sued. It is whether your facts fit the theories courts are allowing, and that is answerable from the account records and a conversation. Evaluations cost nothing, preservation costs nothing, and both protect options that deletion and delay permanently close.
This article is general educational information, not legal advice, and reading it does not create an attorney-client relationship. Allegations described here are drawn from public court filings and are not findings of liability except where a court or jury has so found. Blueshoe's investigation into social media and AI companion harm is open, and submissions are reviewed free of charge and confidentially. If you or someone you love is in crisis, call or text 988, the Suicide & Crisis Lifeline. A licensed Arizona attorney reviews these guides.