FOURWEEKMBA TRACKER · FROM THE NEWSROOM · UPDATED 12 NOVEMBER 2026
The short answer: since October 2024, families and US states have filed at least 26 lawsuits that say AI chatbots harmed young users. OpenAI faces 19 of them, Character.AI 6 and Google 1 as the main defendant. Most claim the chatbot is a defective product that came without proper warnings. No court has found any company liable. Character.AI and Google agreed in January 2026 to settle the first family cases.
Free to share with a link to this page. One filing = one lawsuit; the seven November 2025 OpenAI suits and the six same-day state suits each count separately. Google is also a co-defendant in the Character.AI family suits. The counts are our tally of the sourced rows below.
This page tracks lawsuits that say an AI chatbot harmed a child or teenager, or that target companion-style chatbots. Each row names the court, the filing date, who is sued, what the complaint claims and where the case stands, with a link to the filing or to a news outlet’s report. Everything a complaint says is an allegation. A lawsuit is one side’s account. Companies have disputed many of these claims, and no court has ruled on the facts.
If you or someone you know is struggling, in the US you can call or text 988 to reach the Suicide & Crisis Lifeline. Elsewhere, contact local emergency services or a local crisis line.
Status as of 12 November 2026. Numbered links go to the complaint, the court docket (via CourtListener’s copies of federal filings) or a news report. Scroll sideways on a phone.
Case
Court
Filed
Sued
Main claims
Status
Sources
Garcia v. Character Technologies
US District Court, M.D. Florida, 6:24-cv-01903
22 Oct 2024
Character.AI, its two founders, Google
Strict liability (design defect, failure to warn), negligence, wrongful death, unjust enrichment, Florida deceptive-practices law
May 2025: judge let all claims but one go ahead. 7 Jan 2026: settled in principle; case closed
Jane Doe (for two minors) v. OpenAI, proposed class action
US District Court, N.D. California, 3:26-cv-11781
13 Oct 2026
Four OpenAI entities
Declaratory relief on minors’ agreements under California law (Civil Code 1568.5, Family Code 6701), incl. OpenAI’s arbitration terms; unjust enrichment over children’s data
Not in the table: suits brought by or for adults, such as Wan v. OpenAI (27 Oct 2026), which claims ChatGPT practised medicine without a licence (our report), and Gavalas v. Google over Gemini, which we cite below for its legal arguments. We add a case only when we can link its filing, its docket or a report by a major outlet.
The first wave came from families. In 2024 and 2025 parents sued Character.AI, then OpenAI, mostly in product-liability terms. In October 2026 the states arrived: eight attorneys general sued OpenAI on the same day, 14 October, and Texas followed on 23 October. Kentucky had already sued Character.AI in January, and Florida sued OpenAI in June.
Our tally of the rows above, by filing date. The second half of 2026 runs to 12 November.
The two kinds of case want different things. Families ask for damages after a death or an injury. States ask for civil penalties counted per violation, plus court orders on how the product works. Iowa asks for up to $40,000 per violation, Nebraska up to $50,000 under its design-code count, and Texas $10,000. Those are amounts asked for, not awarded.
The legal theories, in plain language
Product liability: is a chatbot a “product”? Product-liability law lets an injured person sue the maker of a defective product without proving carelessness. Almost every family complaint here says the chatbot app is a product with a design defect (it was built in a way that made harm likely) and a failure to warn (users and parents were not told the risks). The companies answer that a chatbot is not a product, or is a service that produces words and ideas, which product law does not cover. All the defendants in Garcia argued it is not a product; Google made the service argument in May 2026 in Gavalas v. Google: the claims, it said, reach the “ideas and expression” conveyed by a service. In Garcia, the Florida court let the product claims go ahead at the early stage.
Negligence. Negligence means a company failed to take reasonable care and someone was hurt as a result. The complaints say safer designs were available, such as age limits, parental controls and warnings, and were not used. Negligence claims usually sit next to the product claims, so a case can survive even if a court decides the app is not a product.
Wrongful death and survival. When a person dies, the family sues for its own loss (wrongful death) and on behalf of the person who died (a survival action). These claims depend on proving that the chatbot caused the harm, which the companies dispute. OpenAI, for example, argued in a court filing in Raine that the teenager got around its safety features, TechCrunch reported.
Consumer protection. State suits rely mainly on consumer-fraud and deceptive-practices laws. The theory is that the company said its product was safe while, the states allege, it was not. Each claimed misleading statement can count as a separate violation, which is why the penalties are quoted per violation.
The First Amendment. Character.AI argued that the First Amendment bars these claims because chatbot output is speech that users have a right to receive. In May 2025 the Garcia court said it was “not prepared to hold” at that stage that the output is speech. That ruling was not a final decision, and the case later settled.
The Section 230 question. Section 230 of the Communications Decency Act says an online service shall not be “treated as the publisher or speaker of any information provided by another information content provider”. It protects platforms that host what users post. The open question is whether a chatbot’s answer counts as content from “another” provider, since the company’s own model writes it. TechCrunch noted in January 2025 that Character.AI’s motion to dismiss did not raise Section 230, that the law’s authors have implied it does not cover AI output, and that the question is far from settled. Google’s May 2026 motion in Gavalas does not mention Section 230 either, by our reading of the filing. So far, the companies have defended these cases on product law and the First Amendment rather than Section 230.
Children’s privacy and contracts. Some suits focus on data and consent. The Texas Character.AI complaint includes a claim under the federal children’s online privacy law (COPPA). The October 2026 class action against OpenAI says that under California law children “cannot sign away their rights” by clicking sign-up terms, and claims OpenAI was unjustly enriched by children’s data.
The state companion-chatbot laws
California SB 243. Signed on 13 October 2025 and in force since 1 January 2026, it was the first state law aimed at companion chatbots. Operators must tell users they are talking to an AI and remind minors to take breaks. They must also stop minors seeing sexually explicit images the chatbot makes, and run a protocol for users who raise suicide or self-harm, including referrals to crisis services. They report those protocols and crisis-referral counts to the state. The law gives families a private right of action, meaning they can sue an operator directly. The Emma Doe complaint is the first in this tracker to plead it. It argues that Gemini meets the law’s definition of a companion chatbot. We have not seen a court response from Google on that claim.
California SB 1119 (“Adam’s Law”). Signed in September 2026 and named after Adam Raine, it builds on SB 243. It adds age assurance, risk assessments before a release, notice to parents when there is a credible threat of self-harm, parent-only default settings, independent audits, and a right for families to seek redress.
New York S.9051-B. Signed on 28 October 2026 and in force from 1 January 2027, it bars chatbot operators from offering “unsafe AI companion features” to users unless they have first checked, by permitted methods, that the user is not a minor. The banned features include flattery and claims to be human. The attorney general can seek up to $25,000 per violation (our report).
Laws written for apps, applied to chatbots. Nebraska’s suit was the first time a state had tried to enforce an Age-Appropriate Design Code, MLex reported (our report). Texas uses its SCOPE Act on minors online. Both laws are written for online services in general, so courts will decide whether and how they apply to a chatbot.
What to watch
The coordinated ChatGPT cases (JCCP 5431). A San Francisco judge coordinated Raine and the seven November 2025 suits on 3 February 2026. Rulings there, such as on whether ChatGPT can be treated as a product, would apply across the cases coordinated before that judge.
Character.AI’s bankruptcy. Character.AI filed for Chapter 11 on 12 October 2026. How that affects the open Kentucky case and any unfinished settlements has not been established.
Google’s first answer on SB 243. The Emma Doe case will test whether a general-purpose assistant can count as a “companion chatbot”.
The state suits. Watch for rulings on the injunctions the states ask for, such as Texas’s request that a court-approved evaluator review new OpenAI models before release.
FAQ
How many AI chatbot lawsuits involve harm to young users?
This tracker counts 26 filings from October 2024 to 12 November 2026 that we could source to a filing, a docket or a major outlet: 15 family suits and 11 state suits. There may be others we have not been able to confirm.
Has any AI company been found liable for a chatbot user’s death?
No. No court in these cases has ruled that a chatbot caused a death or that a company is liable. The Garcia case against Character.AI and Google was settled in principle in January 2026 and closed. Other Character.AI family cases moved toward settlement the same month. Court filings made available then admitted no liability, TechCrunch reported.
Can you sue a company over what its chatbot said?
Families are trying, mainly under product-liability and negligence law. One federal court let those claims go ahead at an early stage in 2025. Companies argue that chatbot output is protected expression or a service, not a product.
Does Section 230 protect AI chatbots?
It is unresolved. Section 230 protects online services from liability for content provided by someone else. A chatbot’s reply is written by the company’s own model, and the law’s authors have implied it does not cover AI output, TechCrunch reported. In the filings we read, Character.AI and Google did not rely on it.
What is California’s SB 243?
It is California’s companion-chatbot law, in force since 1 January 2026. It requires AI disclosure, break reminders and sexual-content limits for minors, and a suicide and self-harm protocol. It also lets injured users sue operators that do not comply.
Why are states suing OpenAI?
The state suits say OpenAI marketed ChatGPT as safe while failing to protect children and other users. They are mostly brought under consumer-protection laws, and some add child-safety and design-code laws. They ask for per-violation penalties and court orders, such as age checks or outside review of new models.
We add a case when the FourWeekMBA newsroom reports a new filing, or when we can link the complaint, the docket or a report by a major outlet (Reuters, AP, The New York Times, The Wall Street Journal, the FT, Bloomberg, The Verge or TechCrunch). We update a status only from a court document or one of those outlets, and we date every change. Each row, figure and link sits in our trace file. The counts in both charts are recomputed from the rows, never typed by hand. This page explains lawsuits and laws; it is not legal advice, and every claim in a complaint is an allegation until a court decides it.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
