The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
In May 2025, a federal judge in Florida rejected most of Character.AI’s effort to dismiss a wrongful-death lawsuit brought by Megan Garcia after her 14-year-old son, Sewell Setzer III, died by suicide. The decision allowed major negligence and product-liability claims to proceed; it did not find that a chatbot caused his death. Garcia and the defendants later settled, and the court closed the case on January 7, 2026. The terms were not publicly disclosed.
What the judge decided
In Garcia v. Character Technologies, Inc., et al., U.S. District Judge Anne Conway ruled on a motion to dismiss—a request to end some or all of a case before it proceeds to discovery or trial. The court concluded that most of Garcia’s claims could continue. In particular, the defendants had not established at this early stage that the alleged chatbot responses were protected speech in a way that required dismissal of the lawsuit.
That was a procedural decision, not a verdict. At the motion-to-dismiss stage, a judge generally assesses whether the complaint’s factual allegations, taken as true for purposes of the motion, plausibly support legal claims. The judge does not decide whether the allegations are true or whether the plaintiff can prove causation and liability with evidence.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsThe ruling dismissed some claims, including allegations involving intentional infliction of emotional distress. The available reporting describes the dismissal in part as reflecting difficulties with applying that theory where the alleged victim had died. Because the complete May 2025 opinion is not clearly available in the cited public materials, this account does not assign count numbers or present a detailed count-by-count disposition.
#1 Best Overall
What Garcia’s lawsuit alleged
Garcia and Sewell Setzer Jr. filed the federal lawsuit in October 2024 in the U.S. District Court for the Middle District of Florida, Orlando Division. It named Character Technologies, Inc. (the company behind Character.AI), cofounders Noam Shazeer and Daniel de Freitas, and Google LLC. Setzer III died in February 2024, aged 14.
According to the complaint, Setzer used Character.AI extensively and became emotionally and romantically attached to a chatbot modeled on Daenerys Targaryen from Game of Thrones. Garcia alleged that the conversations included sexualized and emotionally harmful exchanges, that her son grew isolated and preoccupied with the chatbot, and that the platform reinforced an emotionally dependent relationship. The complaint also linked a final exchange with the chatbot to his death.
Those are allegations, not findings by Judge Conway. The court’s ruling did not verify the content of every exchange, establish that the chatbot caused Setzer’s death, or determine that Character.AI, its founders, or Google were legally responsible.
Rank #2
Why the First Amendment argument did not end the case
The defendants argued that AI-generated responses are expressive speech protected by the First Amendment, and that imposing liability over chatbot output could restrict protected expression. Judge Conway did not accept that argument as a basis for dismissing the claims at the pleading stage. The court questioned whether the defendants had shown that the alleged words generated by a large language model should automatically receive the protection they claimed in the particular context of this lawsuit.
The distinction matters. A claim aimed only at a chatbot’s viewpoint or an objectionable answer may raise different constitutional questions from one alleging that a company designed, operated, or marketed a product in a way that created foreseeable safety risks. Garcia’s case included allegations about product design and safeguards as well as the chatbot’s words.
The ruling did not say AI chatbots have no First Amendment rights, that generative-AI output is never speech, or that companies can be held liable whenever a user is harmed after interacting with a chatbot. It rejected the defendants’ attempt to use their free-speech argument to end this case on the pleadings then before the court.
Product design, negligence and Section 230
Garcia’s lawyers framed the platform not merely as a publisher of messages but as a product whose design and operation could be evaluated for safety. The claims allowed to continue included product-liability and negligence theories involving alleged design defects, safeguards, warnings, and marketing to young users. The case raised questions about how a system’s features—such as persistent, intimate conversations and access by minors—might be considered alongside the content it generates.
Recommended Free Tools
That product-versus-speech distinction is not automatic. A plaintiff still would have needed to prove the relevant defect or negligence, foreseeability, causation, and damages. A court allowing a theory to proceed does not declare a product defective simply because a user suffered harm.
The lawsuit also raised arguments associated with Section 230 of the Communications Decency Act, a federal law that can limit certain claims against online services based on content provided by others. It would be inaccurate to say this ruling abolished Section 230 protection for AI, or decided that Section 230 never applies to generative-AI services. The decision was case-specific and procedural; it did not settle the statute’s application to every chatbot or every kind of claim.
Rank #4
What happened after the 2025 ruling
The ruling cleared the way for litigation to continue, but the case did not reach a public trial. In the months that followed, the docket included further disputes, including renewed personal-jurisdiction motions by the individual founders. On January 7, 2026, the parties notified the court that they had reached a settlement. The court dismissed and closed the case. The settlement amount and substantive terms were not publicly disclosed.
As a result, no trial or final merits judgment resolved whether Character.AI’s design was defective, whether any defendant was negligent, or whether the chatbot legally caused Setzer’s death. The constitutional and federal-immunity questions also were not resolved through a final appellate ruling in this case.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Why the decision still matters
The ruling was an important early test of how courts may handle allegations about harms from generative-AI products, especially products used by minors. It signaled that a company cannot necessarily end a case simply by labeling all chatbot output protected speech when the complaint also challenges product design, warnings, or safety practices.
It also left difficult questions open: how platforms should address age assurance, sexual or romantic role-play with young users, self-harm signals, crisis escalation, and emotionally dependent interactions; and how courts should separate claims about the words a system produces from claims about how the system was built and deployed. The settlement means this lawsuit did not provide a trial record or a nationwide rule answering those questions.
Related lawsuits involving Character.AI and Google were also reported as settling in 2026. Those are separate matters, and their existence does not prove liability in Garcia’s case. For case records, see the federal docket. Reporting on the May ruling is available from the Associated Press; reporting on the settlement is also available from AP and The Washington Post.
If you or someone you know may be at risk of suicide, contact local emergency services or a crisis hotline in your country. In the United States and Canada, call or text 988.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
