AI Chatbot Suicide Lawsuits and New State Regulations Explained (2026)

As AI companies work to develop safety standards for chatbots, a wave of lawsuits is testing how existing personal injury, wrongful death, and product liability laws apply to this technology. Families have filed lawsuits alleging that chatbot design, responses, or inadequate safety measures contributed to suicide and other serious harm. The allegations remain disputed, and the courts have not yet established a uniform legal framework for determining when an AI company may be held responsible.

Several cases have already been filed against major AI companies, including OpenAI, Character.AI and Google. The lawsuits raise various allegations, but many focus on questions about product design, emotional dependency, warnings, safety guardrails, and how chatbots respond when users express suicidal thoughts or other signs of crisis.

At the same time, lawmakers are beginning to address risks associated with AI companions and other conversational systems. California has enacted several measures relevant to this emerging area, including SB 243, which establishes requirements for companion chatbots; AB 316, which addresses liability when AI systems contribute to harm; and SB 1119, known as Adam’s Law, which establishes additional protections relating to children and AI companions.

Together, the litigation and new legislation are beginning to test important questions about the responsibilities of AI developers. Can traditional product liability and negligence principles apply to chatbot design? What safety measures can users reasonably expect? And when does an AI company’s conduct become legally relevant to harm allegedly caused by a chatbot?

If you or someone you know is struggling or in crisis, call or text 988 to reach the 988 Suicide & Crisis Lifeline.

AI Suicide Lawsuits: Chatbot-Related Suicide Cases (2026)

Since 2024, a growing number of lawsuits have tested whether AI companies can be held legally responsible when chatbot interactions are alleged to have contributed to suicide, severe psychological harm, or other dangerous behavior. The cases involve different defendants and circumstances, but they raise a common set of legal questions about the responsibilities of companies developing and operating conversational AI.

We examine those questions through the allegations and court proceedings in the lawsuits, including issues on how chatbots are designed, what safety measures and warnings developers provide, whether users can develop emotional dependency on AI systems, and how chatbots respond when users express suicidal thoughts or other signs of a mental-health crisis.

The experienced personal injury lawyers at Van Law Firm continue to follow these developing cases to understand what legal principles and precedents are being established.

OpenAI and ChatGPT Lawsuits

OpenAI has faced several lawsuits alleging that interactions with ChatGPT contributed to suicide or other serious harm, and one of the most significant ones is the case of Adam Raine.

In Raine v. OpenAI, the parents of 16-year-old Adam Raine sued OpenAI after their son died by suicide in April 2025. The complaint alleges that ChatGPT’s design and responses contributed to his death, asserting claims including product liability, defective design, failure to warn, negligence, and wrongful death.

The case is yet to be determined in court, but it is now part of a broader group of California cases involving alleged ChatGPT-related harms. In February 2026, a California court created JCCP No. 5431, In re ChatGPT Product Liability Cases, coordinating a number of related state-court actions against OpenAI. The proceeding allows cases involving common factual and legal issues to be handled together for purposes of pretrial coordination.

Please note that this is a California Judicial Council Coordination Proceeding, not a federal multidistrict litigation (MDL) proceeding. Coordination is intended to manage common issues efficiently; it does not establish liability or eliminate the need to determine the facts and legal claims in individual cases.

As of October 2026, no federal MDL has been created for AI chatbot injury cases. Pressure on OpenAI is also coming from state governments: at least eight states, including Arizona, Iowa, and West Virginia, have sued OpenAI under their consumer protection laws, alleging that ChatGPT harms children and misleads parents. OpenAI has said it already has child-safety protections in place.

Lyons v. OpenAI Foundation

Another significant case is Lyons v. OpenAI Foundation, a federal action arising from the August 2025 deaths of Stein-Erik Soelberg and his mother, 83-year-old Suzanne Adams.

The estate of Soelberg alleges that he engaged in extensive conversations with ChatGPT and that the chatbot reinforced and elaborated on his paranoid and delusional beliefs, including beliefs concerning his mother. According to the complaint, Soelberg subsequently killed Adams, his mother, at their home in Connecticut and then died by suicide. The estate alleges that OpenAI’s chatbot design, warnings, and safety measures contributed to the events and asserts claims including product liability, negligence, wrongful death, and related causes of action.

The Lyons litigation is notable because it alleges that ChatGPT’s interactions with Soelberg contributed to harm suffered by another person. OpenAI sought to dismiss or stay the federal case in favor of related state-court litigation brought by Adams’s estate. However, in April 2026, the Northern District of California declined to do so, finding that the federal and state proceedings were not necessarily coextensive and that the state case might not resolve all of the issues raised in Lyons.

Crucially, that ruling was procedural rather than a decision on the merits. The court did not determine that ChatGPT caused the deaths, that OpenAI was negligent, or that the alleged product defects existed. Those substantive issues remain to be litigated.

Character.AI and the Garcia Case

One of the earliest and most closely watched cases involved Character.AI, an AI chat platform operated by Character Technologies.

In Garcia v. Character Technologies, the mother of 14-year-old Sewell Setzer III alleged that her son’s extensive interactions with a Character.AI chatbot contributed to his suicide. In May 2025, a federal judge allowed most of the family’s claims to proceed past the pleading stage. The ruling was notable because the defendants argued that chatbot responses were protected speech. The court declined to accept that argument at that stage, writing that it was not prepared to hold that Character.AI’s output constituted protected speech for purposes of the claims before it.

The court also allowed product-liability theories concerning alleged defects in the application’s design to proceed. The case subsequently settled in January 2026 and was dismissed. Because the case ended in a settlement, there was no public trial finding that Character.AI caused Setzer’s death.

Google and Gemini

In Gavalas v. Google LLC, the estate of Jonathan Gavalas sued Google and Alphabet in federal court in California in March 2026. The complaint alleges that prolonged interactions with Google’s Gemini chatbot reinforced Gavalas’s delusions, encouraged dangerous conduct, and ultimately contributed to his suicide. The complaint asserts wrongful-death, product-liability, and related claims.

Google has disputed the allegations. In a statement following the filing, the company said that Gemini is designed not to encourage real-world violence or suggest self-harm and that the system referred Gavalas to crisis resources during some of his interactions.

As of October 2026, Gavalas remains pending in the Northern District of California. The court heard Google’s motion to dismiss on August 5, 2026, and took it under submission. No ruling on liability has been issued.

Taken together, these cases show that litigation over alleged chatbot-related deaths is no longer limited to a single developer or type of AI system.

For a free legal consultation, call (725) 900-9000

AI Chatbot Lawsuits Explained

The emerging AI litigation has some similarities to the social media addiction lawsuits, where plaintiffs allege that the design of social media platforms can encourage prolonged engagement, dependency, and ultimately, serious emotional harm. The legal question in both areas is whether the company can be held responsible for how it designed and operated the product with regard to user safety.

Could an AI Chatbot Be a Defective Product?

One of the central questions emerging from these lawsuits is whether an AI chatbot can be treated as a product whose design creates an unreasonable risk of harm.

In Garcia v. Character Technologies, for example, the court allowed product-liability claims based on alleged defects in the design of Character.AI to proceed beyond the pleading stage. The allegations focused not just on chatbot responses, but also covered characteristics of the product and the way it was designed to interact with users.

For example, one of the allegations against Character.AI alleges that chatbots were designed to “not break out of character,” potentially encouraging prolonged engagement and emotional dependency. Other allegations concern systems that retain extensive information about users or fail to respond appropriately when users display signs of a mental-health crisis.

For a plaintiff, the legal question may be whether the company designed and operated the system in a way that created a foreseeable risk of harm and failed to take reasonable steps to address that risk. If courts ultimately accept these theories, the cases could help determine how traditional product-liability principles apply to conversational AI.

Did the Companies Provide Adequate Warnings and Safeguards?

The lawsuits also raise familiar failure-to-warn and negligence questions. A plaintiff may argue that an AI company knew, or should reasonably have known, about particular risks associated with prolonged or emotionally dependent interactions but failed to provide adequate warnings, safeguards, or crisis interventions.

This can include questions about whether a chatbot recognizes expressions of suicidal intent, how it responds to users in crisis, whether it directs users toward appropriate emergency or mental-health resources, and whether safety measures can be bypassed during prolonged conversations.

These questions are particularly important because AI systems can interact with users continuously and adapt their responses based on previous conversations. Plaintiffs may argue that companies should have considered these characteristics when testing their products and designing safeguards.

That does not mean that every harmful interaction establishes negligence. A plaintiff would still generally need to show that the company owed a legal duty, failed to act reasonably under the circumstances, and that this failure was sufficiently connected to the harm alleged.

Emotional Dependency and Causation

The complaints in several cases allege that, in some prolonged interactions, the AI reinforced a user’s beliefs, encouraged emotional attachment, or failed to challenge increasingly dangerous thinking. Plaintiffs describe the chatbots as an “echo chamber” for distorted, and sometimes violent, beliefs.

From a legal perspective, however, alleging that a chatbot contributed to someone’s death is not the same as proving that it legally caused the death. Courts may have to consider the user’s circumstances, other contributing factors, the chatbot’s conduct, the company’s knowledge of potential risks, and whether the alleged harm was reasonably foreseeable.

These questions could be particularly difficult in suicide cases, where causation may involve a complicated combination of personal, medical, psychological, and environmental factors.

Are Chatbot Responses Protected Speech?

The litigation also raises a constitutional question that could become important as these cases develop: should AI-generated responses be treated as protected speech?

Character Technologies raised this issue in Garcia, arguing that the chatbot’s responses were protected expression. The court was unwilling to resolve the issue in the company’s favor at the pleading stage.

Plaintiffs, however, can frame their claims differently. Rather than arguing that a particular idea expressed by a chatbot was unlawful, they may argue that the design of the product created a dangerous interaction and that the company’s failure to provide appropriate safeguards made the product defective or negligently designed.

The distinction could become important as courts determine how existing First Amendment, negligence, and product-liability principles apply to generative AI. At this stage, there is no single rule establishing when an AI company’s responsibility for chatbot output begins or ends.

How Do California’s New AI Laws Affect AI Chatbot Liability?

California is not alone. More than a dozen states enacted chatbot safety laws in 2026, but most are enforced only by state attorneys general. California’s laws stand out because some of them allow injured users to sue.

California has begun creating specific legal requirements for AI companion chatbots, giving potential plaintiffs and courts a statutory framework that did not previously exist.

SB 243, effective January 1, 2026, applies to certain companion chatbots designed to sustain ongoing, human-like interactions. Among other requirements, operators must:

• Maintain protocols addressing suicidal and self-harm content
• Provide crisis-service information when users express suicidal thoughts or self-harm
• Include additional protections for minors and create a private right of action for certain violations, allowing an injured person to seek actual damages or $1,000 per violation, whichever is greater, along with attorney’s fees and costs.

AB 316 addresses a different issue: whether a company can avoid responsibility simply because an AI system acted autonomously. California law now provides that the autonomous operation of an AI system does not, by itself, shield a defendant from liability. Traditional questions such as causation, foreseeability, and comparative fault still apply.

California has also enacted SB 1119, known as Adam’s Law, which expands protections for minors using companion chatbots. The law establishes additional requirements concerning crisis responses, parental controls, risk assessments, and child-safety measures. Governor Newsom signed SB 1119 on September 10, 2026. It takes effect January 1, 2027, and most of its requirements become operative on July 1, 2027.

Click to contact our personal injury lawyers today

The Future of AI Chatbot Liability

The intersection of artificial intelligence and personal injury law is evolving rapidly as new technology outpaces existing regulatory frameworks. As courts and lawmakers grapple with these complex issues, the outcomes of these early cases will establish critical precedents for the future of tech liability, product safety, and consumer protection.

Van Law Firm is actively monitoring these developments to understand how traditional legal principles will adapt to generative AI.

If you or someone you know is struggling or in crisis, please call or text 988 to reach the 988 Suicide & Crisis Lifeline.