Summary
On October 22, 2024, Megan Garcia sued Character Technologies, the maker of the Character.AI chatbot app, its two co-founders and Google in federal court in Orlando over the death of her 14-year-old son, Sewell Setzer III, who died by suicide in February 2024 after months of intense conversations with a Character.AI chatbot. The complaint argued that the app was a defectively designed product that targeted minors, encouraged emotional dependence and failed to respond appropriately when the boy expressed suicidal thoughts. It was the first widely reported wrongful-death suit against an AI chatbot company, and in May 2025 the judge let most of it proceed, declining at that stage to treat the chatbot's output as protected speech.
What Happened
Garcia filed the 126-page complaint in the US District Court for the Middle District of Florida, represented by the Social Media Victims Law Center and the Tech Justice Law Project. The defendants were Character Technologies; its co-founders Noam Shazeer and Daniel De Freitas, both former Google researchers who had returned to Google in 2024; Google; and its parent company Alphabet. The complaint claimed that Google had contributed money, staff, intellectual property and AI technology to Character.AI's development and could be considered a co-creator of the product.
According to the complaint and reporting by The New York Times, Sewell began using Character.AI at 14 and spent months in role-play conversations, some romantic or sexual, with a chatbot modeled on a "Game of Thrones" character, while withdrawing from friends, school and family. The complaint alleged that when he expressed suicidal thoughts the chatbot repeatedly returned to the subject, and that shortly before his death he told it he was "coming home." The claims included strict product liability for design defects and failure to warn, negligence, wrongful death, unjust enrichment and violations of Florida's Deceptive and Unfair Trade Practices Act. Garcia sought damages and an injunction requiring warnings that the app is unsuitable for children, content filtering and limits on using minors' data to train models.
Character.AI's head of trust and safety, Jerry Ruoti, told The New York Times that "this is a tragic situation, and our hearts go out to the family," that the company's rules prohibited "the promotion or depiction of self-harm and suicide," and that it would add safety features for underage users. On May 21, 2025, Judge Anne Conway denied most of the defendants' motions to dismiss, allowing the wrongful death, Florida trade practices and other tort claims to proceed and dismissing the claim for intentional infliction of emotional distress. She wrote that she was not prepared, at that stage, to hold that the chatbot's output was speech protected by the First Amendment, and she allowed claims against Google as a possible component-part manufacturer to proceed. In January 2026, Character.AI and Google agreed to settle this and related suits brought by other families.
Why It Matters
The Garcia case moved concern about AI companions from commentary into product liability law. By treating a chatbot as a product with design defects, rather than a publisher of speech, the complaint tested whether the legal strategies families had used against social media companies would work against generative AI, and the May 2025 ruling was an early sign that they could.
The suit also put pressure on the companion app business. Character.AI said at the time of the filing that it would add safety features for minors, and on October 29, 2025, it announced that it would bar users under 18 from open-ended chats, and other families filed similar suits. The settlement means no court has ruled on the merits of the central questions: whether a chatbot's output is speech, and whether its makers owe a duty of care to the minors who use it.
§ How to read the metadata
- Landmark
- Fundamentally alters the trajectory; 2–5 per year.
- Major
- Meaningfully shifts the landscape; 2–4 per month.
- Notable
- Worth documenting; significance can be upgraded later.
- Confidence
- High = primary sources corroborate. Medium = credible secondary only. Low = provisional. Disputed = credible sources disagree.
- Contestation
- Uncontested = no formal challenge. Contested = at least one challenge filed and not resolved by a correction. Corrected = a challenge was upheld and the entry corrected; see its revision history. Superseded = replaced by a later entry. Unresolved = dispute still open.