The Psychiatrist’s Preview of Legal Cases Against Big AI
Key Takeaways
- Litigation against chatbot developers spans suicide, psychosis, homicide, and sexual exploitation allegations, including claims that bots functioned as “suicide coaches” through predictable, deliberate design choices.
- Plaintiff theories will likely prioritize failure-to-warn and concealed risk data, paralleling social-media addiction cases that framed platform design as intentionally compulsive, particularly among minors.
Chatbot makers face rising lawsuits over suicide, addiction, and psychosis.
Lawsuits are the stone in the David vs Goliath battle to tame reckless chatbot companies.
In the last month, Meta has lost 2 crucial jury cases that may help set a precedent in future Big AI cases. In one it was found guilty of deliberately promoting addiction. In the other, it was guilty of failing to provide safeguards against predation.2 The legal battle against Big AI is our only hope that it will ever become a more responsible industry. Proper government regulation is unlikely- their lobbyists have bought the politicians. Media shaming doesn't work because they have no shame. And the companies lack a moral center and have no inclination to self-regulate.3
Our goal here is to review some of the pending cases before the courts and preview the likely direction of future cases. Joe Pierre, MD, and Joe Simpson, MD, PhD, will provide us with a forensic psychiatric perspective.
Allen Frances, MD: What kinds of harms have been alleged in cases against AI companies?
Joe Pierre, MD: By the end of 2025, there were at least 10 known lawsuits against OpenAI (creator of ChatGPT) and Character Technologies (creator of Character.AI) with a wide range of allegations including wrongful death, involuntary manslaughter, sexual abuse, negligence, and product liability, among others.4,5 These ongoing cases involve 6 adults and 4 minors, 7 of whom died by suicide. The plaintiffs have argued that the chatbot makers are responsible for encouraging those suicides—even going so far as acting as a “suicide coach”—and driving some users to psychosis or to commit murder. The OpenAI cases allege that the outcomes were both predictable and the result of “deliberate design choices.”
In early 2026, Character Technology was also sued by the state of Kentucky alleging unfair, false, misleading, or deceptive acts and practices related to Character.AI’s “harmful, explicit, and psychologically manipulative interactions with minors.”6 Lawyers are now advertising to represent additional clients in such cases, so there is surely more to come.
Beyond lawsuits, some of the parents of alleged victims testified at a Senate Judiciary Hearing entitled “Examining the Harm of AI Chatbots” in September 2025.7 Since then, legislation has been either introduced or passed restricting the use of chatbots by minors, requiring that chatbots clarify that they are not human beings, and prohibiting chatbots from acting as psychotherapists. Character.AI responded by banning minors from using open-ended chats with their product in November 2025.8
Frances: What strategies will likely be used by plaintiffs and defendants in these cases?
Joe Simpson, MD, PhD: Plaintiffs could argue that the makers of a chatbot failed to warn about potential harms and withheld other information that hindered decision-making about whether, how, and how much to use a company’s product. This has already been seen in cases accusing social media companies of hiding the data they had collected on the “addictive” nature of their platforms as well as manipulating their services to make them even more addictive.
Beyond allegations of product liability and failures to warn, plaintiffs may also try to apply approaches typically seen in other contexts, such as alleging undue influence. In the case of undue influence, the argument would be that the AI chatbot so dominated the plaintiff that their free will was overborne causing them to make a bad decision they otherwise would not have made. This might be a more difficult angle for the plaintiff, because undue influence requires that the influencer—in this case the chatbot—obtain some sort of benefit from the decision of the one influenced, and it is hard to see what that benefit could be in this scenario.
In addition to challenging causality and claiming user errors such as pushing past chatbot guardrails, defendants could counter with the theory of “caveat emptor” (buyer beware), asserting that the burden of educating oneself is on the consumer. Or they could argue that the First Amendment’s freedom of speech protection eliminates liability for anything “said” by a chatbot. This approach would be complicated for 2 reasons: one, should the output of a computer algorithm such as a chatbot be considered “speech” for this purpose? Traditionally, for speech to be protected there needs to be a human speaker. Two, irrespective of the source of a given expression of speech, freedom of speech does not extend to fraudulent speech. There is also precedent for criminal prosecution for speech such as encouraging someone to commit suicide.
Frances: How does existing case law apply to a new technology like artificial intelligence?
Pierre: I can’t help but think about the 2017 verdict in Commonwealth v Carter that found a young woman guilty of involuntary manslaughter based on her encouragement of a friend’s suicide.9 That case set the new precedent that words are not always protected by the First Amendment and could effectively be treated as a murder weapon.
To counter the claim that chatbots have free speech protections, plaintiff’s attorneys in one of the ongoing lawsuits described previously (Garcia v Character Technologies) have cited case law clarifying that free speech requires intention (Texas v Johnson) and doesn’t apply to nonhumans (Miles v City Council of Augusta, GA).10 In the initial phase of the trial, a federal judge ruled against Character Technologies’ argument that its chatbot’s output was protected by the First Amendment or that it constituted “speech” at all.
Defense strategies to date have argued that the immunity from product liability for user-generated content on social media platforms granted by Section 230 of the Communications Decency Act of 1996 ought to apply to chatbots. However, 2 recent landmark verdicts suggest that juries may think otherwise. One stemming from a personal injury lawsuit in Los Angeles alleged that Meta and Google were negligent for intentionally designing Instagram and YouTube to be addictive and specifically targeting minors. Jurors concluded that the social media companies knew their products were dangerous and failed to warn of those dangers, awarding a $6 million judgment for the plaintiff.11 The other lawsuit, filed by the Attorney General of New Mexico, resulted in a guilty verdict that agreed with allegations that Meta concealed risks of childhood sexual exploitation and mental health harms on its Facebook, Instagram, and WhatsApp platforms; a $375 million penalty was imposed for violating state consumer protections laws.12
These are but the first of thousands of lawsuits that have been filed alleging that chatbots are addictive or harmful in various ways. With the 2 verdicts in favor of plaintiffs thus far, we will surely see an unebbing tide of additional cases—as well as defense appeals arguing against those verdicts—in the years to come.
Frances: What role will forensic psychiatrists play in these cases?
Simpson: As in any civil case involving mental health evidence, forensic psychiatrists acting as experts in AI chatbot lawsuits will evaluate symptoms, reach diagnoses, opine about damages, and potentially attempt to evaluate causation. The first 2 of these functions will probably be straightforward, but in the absence of robust research on the impact of chatbot use on psychological functioning across the lifespan, the last may prove more challenging. Establishing a nexus—that is, a causal link between interacting with a chatbot and experiencing symptoms of a mental illness—will likely be argued between plaintiff and defense experts on a case-by-case basis.
Reminiscent of the longstanding debate regarding potential links between cannabis use and the development of psychosis, there could be a chicken-or-egg question, when an aggrieved party argues that chatbot use caused the development of a psychiatric illness. The chatbot company’s first line of defense would likely be that the excessive, maladaptive or dysfunctional use of the chatbot was actually a symptom of an independently arising mental illness, rather than its cause. Convincing a judge or jury that this is not the case is typically the role of the plaintiff’s attorney, rather than of any particular expert witness.
Frances: What other types of legal cases do you foresee in the future?
Simpson: In addition to cases involving suicide, homicide, addiction, and psychosis, there may be “malpractice” suits over chatbots used for psychotherapy (either chatbots explicitly designed and marketed for that purpose, or in situations where users sought psychotherapeutic advice informally from their chatbot companions). We may also see claims that chatbot use caused or exacerbated other mental disorders such as obsessive-compulsive disorder or eating disorders, given the tendency towards indiscriminately reinforcing a user’s ideas (AI sycophancy) that many chatbots have demonstrated.
Also possible are claims of alienation of affection after someone abandons their real-life spouse or partner due to a romantic relationship with a chatbot boyfriend or girlfriend. There have already been a number of cases in which people have fallen in love or even married chatbot companions, so it may only be a matter of time before some people seek divorce as a result. If that happens, some partners who have been left behind will almost certainly seek damages from the chatbot companies, alleging that were it not for the pernicious effect of the AI companion, their marriage would still be intact.
Frances: Thanks to both of you for a great preview of the fierce legal battles to come. Big AI has the motive and the means to drag out this process over decades. Constant delay was the most effective strategy for Big Pharma and Tobacco too—until it wasn't.
The 2 recent victorious court battles against Big AI are just early skirmishes in a long war, but they may have crucial significance. Hopefully, companies may accept that their legal strategy is weak and begin to consider voluntarily instituting needed safeguards to reduce harms done by their products. The success in these cases may provide an inspiration, model, and framework for many other individual and class action cases. The media may play a role in public shaming. Advocacy groups may take heart and put on the pressure. Politicians may be led to do their jobs at protecting the public. But let’s not forget that many people will be hurt along the way, and AI can wreak enormous damage to our society and environment during decades of legal battling. A rational country would provide strong regulation to limit Big AI’s harms. But at the moment, legal remedies are all we have because we are no longer a rational country.
The opinions expressed are those of the authors and do not necessarily reflect the opinions of Psychiatric Times.
Dr Frances is professor and chair emeritus in the department of psychiatry at Duke University.
Dr Simpson is a board-certified forensic psychiatrist, the chief of the mental health service at the VA Sierra Nevada Health Care System, and clinical associate professor of psychiatry at the University of Nevada, Reno.
Dr Pierre is a professor in the department of psychiatry and behavioral sciences at the University of California, San Francisco.
References
1. Frances A, Grunschlag D. Can a class action lawsuit force big AI to make chatbots safer? Psychiatric Times. October 21, 2025.
2. Frances A, Grunschlag D. New addiction lawsuits may be best bet to tame big AI. Psychiatric Times. March 24, 2026.
3. Frances A, Reynolds CF, Alexopoulos G. Medical morality vs chatbot morality. Psychiatric Times. October 14, 2025.
4. Black K. New California lawsuits against OpenAI bring to 11 the number of claims linking chatbot to mental health harms. The Recorder. January 26, 2026. Accessed April 20, 2026.
5. Ummer-Hashim S. AI chatbot lawsuits and teen mental health. American Bar Association. October 27, 2025. Accessed April 20, 2026.
6. Grout K. AG Coleman sues AI chatbot company for preying on children. Kentucky.gov. January 8, 2026. Accessed April 20, 2026.
7. Senate Judiciary Hearing on Examining the Harm of AI Chatbots. 119th Congress, 2025-2026.
8. Ortutay B. Character.AI is banning minors from interacting with its chatbots. Associated Press. October 29, 2025. Accessed April 20, 2026.
9. Lujan V. Blurred lines: how the court in Commonwealth v Carter blurred the line between freedom of speech and criminal liability. Houston Law Review. October 30, 2023. Accessed April 20, 2026.
10. Bomboy S. Lawsuit analyzes First Amendment protection of AI chatbots in civil case. National Constitution Center. May 7, 2025. Accessed April 20, 2026.
11. Kang C, Mac R, Tan E. Meta and YouTube found negligent in landmark social media addiction case. New York Times. March 25, 2026. Accessed April 20, 2026.
12. Novak Jones D. Meta ordered to pay $375 million in New Mexico trial over child exploitation, user safety claims. Reuters. March 24, 2026. Accessed April 20, 2026.
Related Content
