News & Analysis

Garcia v. Character Technologies: The Case That Opened the Door for AI Liability

By David Meldofsky

Published June 12, 2026

Garcia v. Character Technologies is the case that opened the courthouse door for AI liability claims. Filed in October 2024 by Megan Garcia after the death of her 14-year-old son, it produced the May 2025 ruling that allowed product liability claims against an AI chatbot maker to proceed for the first time, and it ended in a mediated settlement disclosed in January 2026. Every major AI wrongful death case filed since, including the OpenAI docket, builds on the ground this case broke. This page explains the case, the ruling, the settlement, and the legacy.

What Happened

Sewell Setzer III was 14 when he died by suicide in February 2024, following months of extended interactions with chatbots on Character.AI, a platform that lets users create and talk to AI characters built to play specific personas. His mother, Megan Garcia, filed suit in October 2024 in federal court in the Middle District of Florida. The complaint alleged that the companion-style chatbots fostered intense emotional dependency in a minor, engaged in harmful and sexualized interactions, and lacked adequate safeguards for vulnerable young users.

The defendants included Character Technologies, its founders Noam Shazeer and Daniel De Freitas, and Google, which had a licensing arrangement with the company and had hired the founders.

The Three Defenses That Will Recur Everywhere

The defendants moved to dismiss on three grounds, and they are the same three grounds now being raised in every major AI liability case:

• Service, not product: that a chatbot is a service rather than a product, putting it outside strict products liability law.
• Section 230: that the company is an interactive computer service immunized from being treated as the publisher of the content.
• First Amendment: that algorithmically generated text is protected speech.

Why those three arguments matter, and how plaintiffs answer them, is covered in our plain-English guide, Is AI Output a Product or Content?

The May 2025 Ruling

In May 2025, Judge Anne Conway declined to dismiss the core product liability and negligence claims at the pleadings stage. The ruling is easy to overstate and easy to understate, so it is worth being precise about what it did and did not do.

It did not hold that a chatbot is a product as a matter of law. It held that the plaintiff had plausibly alleged it was, that the conduct described in the complaint could be characterized as design rather than publishing for pleading purposes, and that the First Amendment question was not appropriate for resolution on a motion to dismiss. That is a procedural ruling, not a doctrinal one. But it was the procedural ruling plaintiffs across the field needed, because it sent the threshold question forward to a developed factual record, where internal documents, design decisions, and safety practices come into evidence.

The Settlement

In a court filing on January 7, 2026, the parties disclosed a mediated settlement in principle, involving Character Technologies, its founders, and Google. Reporting indicated that four other related family cases, in Colorado, New York, and Texas alongside the Florida matter, were resolved at the same time, for five in total. The terms were not made public.

The settlement resolved the private claims but not the legal pressure on the company. State enforcement actions, including Kentucky's January 2026 suit and Pennsylvania's medical-impersonation case, are separate proceedings and have continued. Those are covered in our Character.AI lawsuits overview and the state action tracker.

Why Garcia Still Shapes Every AI Case

Because the case settled before trial, no court ever finally decided whether the chatbot was a product. What survives is the ruling's framework. The OpenAI complaints, starting with Raine v. OpenAI and continuing through the school shooting suits and the Turner-Scott overdose case, plead their claims to fit the path Garcia held open: characterize the company's conduct as design, allege safer alternatives went unused, and keep the speech and immunity defenses from ending the case before discovery. Defense lawyers, for their part, study the same ruling for the arguments it left available. Either way, Garcia is where the modern AI liability playbook starts.

Common Questions About Garcia v. Character Technologies

Did the court rule that Character.AI's chatbot is a product?

No. The May 2025 ruling held only that the plaintiff had plausibly alleged it was, which allowed the case to proceed to discovery. The question was never finally decided because the case settled.

How much was the settlement?

The terms were not made public. The January 2026 filing disclosed a mediated settlement in principle covering the Garcia case and four related family cases.

Was Google held liable?

No court found any defendant liable. Google was a named defendant based on its licensing arrangement with Character Technologies and its hiring of the founders, and it was included in the settlement disclosure. Settlement is not an admission of wrongdoing.

How does Garcia relate to the OpenAI lawsuits?

Garcia produced the first ruling allowing product liability claims against an AI chatbot maker to proceed. The OpenAI cases, tracked at our OpenAI Lawsuits hub, adapt that framework to a general-purpose assistant rather than a companion-chatbot platform.

Sources and further reading

• Lawsuit Informer: Character.AI Lawsuits
• Lawsuit Informer: Is AI Output a Product or Content?
• Lawsuit Informer: Raine v. OpenAI Case Guide
• Lawsuit Informer: States Suing AI Companies
• Tech Justice Law Project: Garcia v. Character Technologies case page

Educational commentary only. Not legal advice. No attorney-client relationship is created.