ChatGPT sits between product and service in United States courts. OpenAI calls it an interactive service outside traditional strict product-liability law, while plaintiffs sue it as a defective product with missing safeguards. Strict product liability means a maker pays for a defective item even without proven carelessness. That label decides whether families can use strict-liability theories or must prove negligence and faulty warnings.
Official resources:
- Read the official notice from Hunton — Use this primary source to verify the official announcement.
- Read the official notice from Addingtonlaw — Use this primary source to verify the official announcement.
Table of Contents
- Why the label controls the legal path
- What the California group cases allege
- What early chatbot rulings signal
- Why Europe gives a clearer product answer
- What families and potential claimants should save
Why the label controls the legal path
A product case can allege defective design and failure to warn. A service case usually requires proof that the provider acted without reasonable care. Lawyer Monthly describes OpenAI's position as an interactive service rather than a product in its October product-versus-service analysis.
Plaintiffs answer that sycophantic responses and weak safeguards are design defects. Consumer-protection claims add another path. They focus on deceptive conduct and inadequate safety testing rather than physical defect alone.
What the California group cases allege
Hunton Andrews Kurth reports that twelve ChatGPT cases were coordinated in February 2026 in San Francisco County Superior Court as In re: ChatGPT Prod. Liab. Cases, JCCP No. 5431, in its California consolidation update.
The suits allege sycophantic design and inadequate testing caused psychological harm and suicides. The same Hunton Andrews Kurth update says plaintiffs invoke strict product liability, negligence, failure to warn, and state consumer-protection claims. Those are the traditional theories used in other United States large-language-model cases. In plain terms, design defect means the system encouraged harmful reliance. Failure to warn means vulnerable users got no effective caution about hallucinations, emotional dependence, or self-harm risk.
What early chatbot rulings signal
A federal court refused to dismiss strict-liability, negligence, and Florida deceptive-practices claims in Garcia v. Character Technologies, 785 F. Supp. 3d 1157, according to Hunton Andrews Kurth. That case involved a teen suicide allegedly linked to a chatbot.
📨 Get Free Mass Tort Guides Alerts
Free · No spam · Unsubscribe anytime
Addington Law summarizes a defense win in Walters v. OpenAI, where a Georgia court granted summary judgment in May 2025 in its early defamation review. ChatGPT had falsely linked radio host Mark Walters to embezzlement, but the court found no negligence, actual malice, or concrete damages. The limits also appear in violent-act claims. Best Lawyers reports a wrongful-death suit over the April 17 2025 Florida State University shooting alleges the shooter asked ChatGPT what weapon to use, when the student union was most crowded, and which target brought more exposure. False or disturbing output alone does not create liability without fault, causation, and real injury.
Why Europe gives a clearer product answer
Bird & Bird explains that the European Union revised Product Liability Directive 2024/2853 takes effect from December 9 2026 and expressly treats AI software as a product in its directive briefing. Defective AI can trigger uncapped strict liability for injury, property loss, or private-data destruction.
The fault-based companion failed. LegalClarity notes the separate AI Liability Directive proposal from September 2022 was formally withdrawn in 2025 after no agreement. For United States readers, the contrast matters because European product status does not decide American service-versus-product fights.
What families and potential claimants should save
Coordination is not compensation. Each California plaintiff must still prove defect, warning failure, causation, and harm under the chosen theory.
Bring those materials to a personal-injury or consumer-protection lawyer promptly. Ask how the JCCP coordination, filing deadlines, and defamation damage rules affect an individual claim.
- full chat exports with dates, prompts, and account identifiers
- medical, counseling, school, and work records showing timeline and harm
- screenshots, witness names, and proof of reliance on the output
- purchase, subscription, and parental-control records where relevant
You Might Also Like
- What Evidence Proves Negligence in a Product Liability Lawsuit?
- Monsanto Roundup Lawsuit: Liability Theories Behind the October Settlement
- How to Choose a Law Firm for a Product Liability Claim