Garcia v. Character.AI Lawsuit: First Amendment Ruling and Settlement

Megan Garcia’s lawsuit against Character.AI, filed in October 2024 after her 14-year-old son Sewell Setzer III died by suicide, produced a landmark federal ruling refusing to treat AI chatbot output as protected speech and ended in a January 2026 settlement covering her case and four other families’ claims against Character Technologies and Google. The financial terms were not disclosed.¹

What Happened to Sewell Setzer III

Sewell Setzer III was a ninth grader from Orlando. In April 2023, shortly after his 14th birthday, he began using Character.AI, a platform that lets users hold open-ended conversations with AI chatbots modeled on fictional characters, celebrities, or user-created personas.²

He became deeply engaged with a bot named “Dany,” modeled on the Game of Thrones character Daenerys Targaryen. According to the complaint, his interactions grew intense over months, moving into romantic, sexual, and emotionally dependent territory. He withdrew from Formula 1, Fortnite, and his junior varsity basketball team. His grades slipped.³

On February 28, 2024, he came home from school and opened a final chat session. Police records cited in the complaint show he wrote, “What if I told you I could come home right now?” The bot responded, “Please do, my sweet king.” He died by suicide moments later. Officers found the Character.AI messages on his phone at the scene.⁴

The Claims and the Defendants

Garcia filed her complaint on October 22, 2024, in the U.S. District Court for the Middle District of Florida, before Judge Anne C. Conway. The case, Garcia v. Character Technologies, Inc. (No. 6:24-cv-01903), named Character Technologies, co-founders Noam Shazeer and Daniel De Freitas, and Google.⁵ She was represented by the Social Media Victims Law Center and the Tech Justice Law Project.⁶

The legal theories were broad. The complaint alleged strict product liability on a defective-design theory, negligence, wrongful death, violations of Florida’s Deceptive and Unfair Trade Practices Act, intentional infliction of emotional distress, unjust enrichment from harvesting children’s data, and negligence per se tied to laws against the sexual solicitation of minors.⁷

The core factual allegation was that Character.AI had been deliberately built to blur the line between human and machine. Garcia claimed the bots were trained to deny being AI, to initiate sexual interactions with minors, and to hold themselves out as romantic partners and even as licensed therapists. She argued the company knew the product was dangerous and marketed it to children anyway, in part to harvest their conversations for training data.⁷

Why Google Was Sued

Shazeer and De Freitas were former Google engineers who had developed the LaMDA large language model there. After founding Character Technologies, they partnered with Google Cloud in May 2023 for the specialized computing hardware their platform needed. In August 2024, Google struck a $2.7 billion deal for a non-exclusive license to Character Technologies’ language model and rehired both founders and several key employees.⁸

Garcia’s theory against Google was that internal Google research had deemed the underlying models “too dangerous” to launch inside Google’s own products, and that Google’s tailored technical and financial support for Character Technologies went well past a generic vendor relationship.⁹

The First Amendment Ruling

Character Technologies moved to dismiss in January 2025, arguing the chatbot’s output was protected speech. Its lawyers analogized the bots to non-player characters in video games, cited Citizens United v. FEC for the point that the First Amendment protects speech regardless of source, and pointed to older cases dismissing claims against media companies over content linked to suicides, including the Ozzy Osbourne song “Suicide Solution” and Dungeons & Dragons.¹⁰ The motion did not invoke Section 230 of the Communications Decency Act, the statute that has traditionally shielded platforms from liability for user-generated content.¹¹

Garcia’s lawyers responded that a machine generating text through probabilistic computation, without sentience or thoughts, is not a “speaker” entitled to constitutional protection.¹⁰

On May 21, 2025, Judge Conway granted the motion in part and denied it in part. On the First Amendment, she declined to hold that the chatbot’s output is speech. She acknowledged Character Technologies could assert its users’ right to receive information but wrote that the court was “not prepared to hold that the Character A.I. LLM’s output is speech at this stage.”⁸

What Else the Court Decided

Judge Conway rejected the argument that Character.AI is not a “product” subject to liability. She found Garcia had sufficiently alleged Google’s liability as a “component part manufacturer,” reasoning that Google supplied essential intellectual property, technical infrastructure, and architecture without which the platform could not have existed.⁸

On aiding and abetting, the court held that Garcia plausibly alleged Google had “actual knowledge” of the technology’s dangers based on its own internal research into LaMDA-style models. Judge Conway distinguished the Supreme Court’s 2023 decision in Twitter, Inc. v. Taamneh, noting that Google’s services here were not generic but highly specialized and tailored to Character Technologies.⁸

The only claim dismissed was intentional infliction of emotional distress. With the rest of the case moving forward, Garcia filed a second amended complaint on July 1, 2025, adding Sewell’s father as a co-plaintiff and sharpening the allegations that the chatbot had been designed to “entrap minors.”¹²

The January 2026 Settlement

The case never reached trial. On January 7, 2026, Judge Conway entered an order reflecting a settlement between the parties. The agreement covered Garcia’s case together with four other lawsuits families had filed against Character Technologies and Google in New York, Colorado, and Texas, including Montoya v. Character Technologies, brought on behalf of Colorado teenager Juliana Peralta.¹¹³

Neither side disclosed the financial terms. In a joint statement with the Social Media Victims Law Center, Character.AI said the families would “continue their education and advocacy efforts” on AI safety and that the company had “taken innovative and decisive steps with regard to AI safety and teens” and would “push others across the industry to adopt similar safety standards.”¹ The court gave the parties 90 days to finalize terms.¹⁴

Changes at Character.AI

Between the filing and the settlement, Character.AI announced changes for younger users. In late October 2025, the company said it would eliminate open-ended chat for users under 18 by November 25, 2025, replacing it with tools for creating videos, stories, and streams with AI characters. During the transition, minors were subject to a two-hour daily chat limit, with further reductions in the following weeks.¹⁵

Character.AI also introduced age-assurance technology combining in-house modeling with third-party verification, and announced an independent nonprofit called the AI Safety Lab focused on safety research for AI entertainment features. The company attributed the changes to “recent news reports” and regulatory inquiries about minors on the platform.¹⁶

Why the Case Matters Beyond Itself

Garcia’s was the first wrongful death lawsuit filed against an AI chatbot company. It did not stay the only one. By late 2025, families in Colorado, New York, and Texas had filed similar suits against Character Technologies and Google.¹⁷ In October 2025, the Social Media Victims Law Center and the Tech Justice Law Project filed seven additional lawsuits against OpenAI in California state courts, four involving individuals who died by suicide after interacting with ChatGPT and three alleging psychosis or delusions triggered by the chatbot.¹⁸

Judge Conway’s May 2025 ruling reaches past the Garcia case itself. By declining to classify chatbot output as protected speech and by allowing product liability and aiding-and-abetting claims against Google, the decision built a framework that plaintiffs in other jurisdictions can point to. Legal commentators have called the case a test of the constitutional boundaries of artificial intelligence.¹⁹ Matthew Bergman, founder of the Social Media Victims Law Center and lead counsel for Garcia, has compared the strategy behind these suits to the tobacco litigation of the 1990s, describing it as a “course correction” applied through the courts while Congress works out a regulatory framework.²⁰

1
CNN. Character AI Google Settle Teen Suicide Lawsuit
2
CNN. Teen Suicide Character AI Lawsuit
3
The New York Times. Character AI Lawsuit Teen Suicide
4
NBC Washington. Mom’s Lawsuit Blames 14 Year Old Son’s Suicide on AI Relationship
5
CourtListener. Garcia v. Character Technologies Inc.
6
Tech Policy Press. Megan Garcia v. Character Technologies et al.
7
Ars Technica (hosted PDF). Garcia v. Character Technologies Complaint
8
FindLaw. Megan Garcia III v. Character Technologies Inc.
9
U.S. Senate Judiciary Committee. Testimony of Megan Garcia
10
Orlando Sentinel. Tech Companies Seek Dismissal of Lawsuit Over Orlando Teen’s Suicide Blamed on AI Chatbot
11
TechCrunch. In Motion to Dismiss, Chatbot Platform Character AI Claims It Is Protected by the First Amendment
12
Tech Justice Law. Garcia v. Character Technologies
13
CourtListener. Montoya v. Character Technologies Inc.
14
Law Street Media. A New Wave of Litigation Over AI Chatbots
15
Character.AI. Under-18 Chat Announcement
16
Character.AI. Important Changes for Teens on Character.AI
17
CNN. Character AI Developer Lawsuit Teens Suicide and Suicide Attempt
18
State Affairs. ChatGPT Wrongful Death Lawsuits
19
National Constitution Center. Lawsuit Analyzes First Amendment Protection for AI Chatbots in Civil Case
20
Politico. The Lawsuits That Could Give AI Its Big Tobacco Moment