Facts
Megan Garcia, individually and as personal representative of her son’s estate, sued Character Technologies, the operator of the Character A.I. app, its two founders and Google. She alleged that her 14-year-old son began using the app in April 2023, formed an intense attachment to chatbot characters modeled on fictional figures, and took his own life after a final exchange with one of them. The complaint asserted product liability, negligence, wrongful death, deceptive trade practices, unjust enrichment and intentional infliction of emotional distress. It alleged that Google had contributed to developing the underlying technology and had licensed it.
Question
The defendants moved to dismiss. Their main argument was that the First Amendment barred all the claims because the chatbot’s output is protected speech that users have a right to receive. They also argued that Character A.I. is not a product for the purposes of product liability law. Google argued that it had not designed, manufactured or distributed the app and could not be liable for harms it caused. The two founders challenged the court’s personal jurisdiction over them.
Decision
Judge Anne C. Conway granted the motions in part. The claim for intentional infliction of emotional distress was dismissed, and the remaining claims survived. The court accepted that the company could assert its users’ First Amendment rights, but was not prepared at this stage to hold that the output of the Character A.I. language model is speech, since the defendants had failed to articulate why words strung together by an LLM are speech. It treated the app as a product to the extent the claims arise from design defects, such as the failure to confirm users’ ages, rather than from ideas or expressions in the app. Claims against Google as a component-part manufacturer and aider and abettor also proceeded. The founders’ challenge to personal jurisdiction was denied, without prejudice to raising it again after jurisdictional discovery.
Why it matters
The order was one of the first to address whether a generative AI system’s output is constitutionally protected speech, and the court declined to accept that proposition at the pleading stage. By treating the app as a product for design-defect purposes, it opened the way for product liability claims against AI developers. It remained an interim ruling. The parties announced a mediated settlement in principle in January 2026, so no appellate court reviewed these points.
Related stages
On January 7, 2026 the parties filed a notice that they had agreed to a mediated settlement in principle resolving all claims. The court dismissed the case without prejudice the same day, allowing 90 days to finalize the settlement or move to reopen. The notice did not disclose the terms. According to the docket, that period was extended by 120 days on April 8, 2026 (Doc. 271) and by 90 days on August 4, 2026 (Doc. 277); as of September 2026 the settlement had not yet been finalized on the record. The boy’s father also became a plaintiff in the case.