# AI Liability: Who Is Liable When an AI System Causes Harm, Under EU Product Liability Law, US Courts and State Statutes

Source: https://superintelligencecompliance.ai/guides/ai-liability  
Publisher: ZSKFL Management  
Last updated: 2026-11-16. Not legal advice.

> EU product liability reaches defective AI software placed on the market or put into service on or after Dec 9, 2026. US cases, state defenses, Section 230.

## At a glance

- No US statute and no EU act sets a general liability rule written for AI. Liability comes from existing product liability, negligence, discrimination, consumer protection and professional rules, applied to the companies and people who build, sell and use the system.
- Under the revised EU Product Liability Directive, Directive (EU) 2024/2853, “product” includes software, and recital 13 says AI system providers within the meaning of the AI Act should be treated as manufacturers. It applies to products placed on the market or put into service on or after December 9, 2026, the transposition deadline.
- The Commission has withdrawn the proposed EU AI Liability Directive. The withdrawal was published in the Official Journal on October 6, 2025 (C/2025/5423).
- California (Civil Code 1714.46, from January 1, 2026) bars a defendant who developed, modified, or used AI alleged to have caused a harm to the plaintiff from asserting that the AI autonomously caused the harm to the plaintiff; the defendant may still present any other affirmative defense. Utah provides that it is not a defense to a violation of a statute its Division of Consumer Protection administers and enforces that generative AI made the violative statement, undertook the violative act, or was used in furtherance of the violation.

## Who is liable when an AI system causes harm?

The companies and people behind it, under the law that already governs the harm. Depending on the claim, that can be the manufacturer or developer of a defective product, the business that deploys the system, for what it says and decides, or, in some cases, a vendor that performs a function on its customer’s behalf. No court or statute on this page treats the AI system itself as the responsible party. In the EU, the revised Product Liability Directive applies no-fault liability to defective software, including AI, placed on the market or put into service on or after December 9, 2026.

The *Garcia* and *Mobley* liability rulings below came on motions to dismiss. A ruling that a claim may proceed is not a finding of liability.

**AI liability theories and rules, as of November 15, 2026 (not a complete list)**

| Theory or law | Jurisdiction | Who can be liable | Key requirement | Status or date |
|---|---|---|---|---|
| Product Liability Directive (EU) 2024/2853 | EU, through national transposing law | Manufacturers of a defective product or component, including software and AI system providers; for manufacturers outside the Union, importers, authorised representatives and, where there is neither in the Union, fulfilment service providers; where no liable operator established in the Union can be identified, distributors, and some online platforms, that fail to identify one within one month of a request | A defective product causes death or personal injury (including medically recognised damage to psychological health), damage to property other than the defective product itself, a product damaged by a defective component integrated into it by or under the control of its manufacturer, and property used exclusively for professional purposes, or destruction or corruption of data not used for professional purposes; no fault needed | Transposition due December 9, 2026; applies to products placed on the market or put into service on or after that date (Article 2(1), as corrected May 7, 2026) |
| Proposed AI Liability Directive, COM(2022) 496 | EU | None; it was a proposal | None | Withdrawal published in the Official Journal October 6, 2025 |
| Product liability and negligence | US states (common law) | Developer of an AI app; a supplier pleaded as a component part manufacturer | In *Garcia v. Character Technologies*, product liability claims could proceed so far as they arose from defects in the app rather than ideas or expressions within it | Motions to dismiss denied except as to the intentional infliction of emotional distress claim, May 21, 2025 (M.D. Fla.); dismissed with prejudice after settlement, October 19, 2026 |
| Agent liability under anti-discrimination law | US federal; California (FEHA) | An AI hiring vendor alleged to act on the employer’s behalf, and the employer | Plausible allegation that the vendor performs a hiring function the employer delegated to it | *Mobley v. Workday* (N.D. Cal.): pending; class certification hearing set for March 9, 2027 |
| Negligent misrepresentation by a chatbot | Canada (British Columbia) | The business whose website chatbot gave the information | Duty of care; an untrue, inaccurate or misleading representation made negligently; reasonable reliance; resulting damages | *Moffatt v. Air Canada*, 2024 BCCRT 149, small claims decision of the Civil Resolution Tribunal, February 14, 2024 |
| California Civil Code 1714.46 (AB 316, Chapter 672, Statutes of 2025) | California | A defendant who developed, modified, or used AI alleged to have caused a harm to the plaintiff | It shall not be a defense, and the defendant may not assert, that the AI autonomously caused the harm to the plaintiff; the defendant may still present any other affirmative defense, including evidence relevant to causation or foreseeability, and other evidence relevant to the comparative fault of any other person or entity | Approved October 13, 2025; in force January 1, 2026 |
| Utah Code 13-77-102 (enacted by SB 226 of 2025 as 13-75-102) | Utah | A person who violates a statute administered and enforced by the Division of Consumer Protection | It is not a defense that generative AI made the violative statement, undertook the violative act, or was used in furtherance of the violation | Effective May 7, 2025 |
| Texas HB 149, Business & Commerce Code 552.105(e) | Texas | Defendants in an Attorney General action under Chapter 552 | A defendant may not be found liable if another person used its AI system in a prohibited manner, or if it discovered the violation through feedback from a developer, deployer, or other person; through testing, including adversarial or red-team testing; by following guidelines set by applicable state agencies; or through an internal review process, if it substantially complies with the most recent version of the NIST Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile or another nationally or internationally recognized AI risk management framework | Effective January 1, 2026; Attorney General enforcement, no private right of action |
| Colorado SB 26-189, C.R.S. 6-1-1707 and 6-1-1709 | Colorado | Developers and deployers of covered automated decision-making technology (ADMT) | No new private right of action. In an existing unlawful-discrimination suit arising from a decision that covered ADMT materially influenced, fault is allocated between developers and deployers by relative fault, and a contract clause indemnifying a developer or deployer for its own violations of Colorado anti-discrimination law is void. No framework-based defense; compliance with Part 17 is not a defense to, and does not excuse, noncompliance with any other law | Consequential decisions made on or after January 1, 2027; Attorney General enforcement limited by *X.AI LLC v. Weiser* |
| California SB 243 (Chapter 677, Statutes of 2025) | California | Operators of companion chatbot platforms | Private right of action: injunctive relief, the greater of actual damages or $1,000 per violation, and reasonable attorney’s fees and costs | January 1, 2026 |
| Section 230, 47 U.S.C. 230(c)(1) | US federal | A shield, not a source of liability, for providers and users of interactive computer services | No treatment as publisher or speaker of information provided by another information content provider; applied in *Kohls v. Bonta* to platforms, not to a model’s own output | AB 2655 judgment August 20, 2025; state appealed September 19, 2025; no appellate decision found as of November 16, 2026 |
| Rules of professional conduct; court sanctions | US states and courts | The lawyer who uses the tool or signs the filing | Competence, confidentiality, communication, candor, supervision and fee rules applied to generative AI (ABA Formal Opinion 512) | Opinion issued July 29, 2024; not binding by itself |

## Is AI software a “product” under EU product liability law?

Yes, for products placed on the market or put into service on or after December 9, 2026 (Article 2(1), as corrected in the Official Journal on May 7, 2026); Directive 85/374/EEC still governs earlier products (Article 21). Article 4(1) of Directive (EU) 2024/2853 defines “product” to include software. Recital 13 names AI systems as software, says software is a product however it is supplied, including as software-as-a-service, and says a developer or producer of software, including AI system providers within the meaning of the AI Act, should be treated as a manufacturer. It also says information is not a product, so the rules do not apply to the content of digital files, such as media files or e-books, or to the mere source code of software. Free and open-source software developed or supplied outside the course of a commercial activity is excluded (Article 2(2)).

The Directive does not affect national contractual liability or non-contractual liability on grounds other than defectiveness (Article 2(4)(b)), so fault-based claims under national tort law remain available alongside it.

**What Directive (EU) 2024/2853 changes for software and AI**

| Article | Rule |
|---|---|
| 7(1), 7(2)(c) and (f) | A product is defective where it does not provide the safety that a person is entitled to expect or that is required under Union or national law. All circumstances count, including the effect on the product of any ability to continue to learn or acquire new features after it is placed on the market or put into service, and relevant product safety requirements, including safety-relevant cybersecurity requirements. |
| 8(1) and (2) | Liable: the manufacturer of the defective product; the manufacturer of a defective component integrated into or inter-connected with a product within the manufacturer’s control that caused that product to be defective; for a manufacturer established outside the Union, the importer, the authorised representative and, where there is neither in the Union, the fulfilment service provider. A person that substantially modifies a product outside the manufacturer’s control and then makes it available or puts it into service is treated as its manufacturer. |
| 9(1), (3) to (5) | A claimant who has presented facts and evidence sufficient to support the plausibility of the claim can obtain an order requiring the defendant to disclose relevant evidence at its disposal, limited to what is necessary and proportionate, with measures to protect trade secrets. |
| 10(2) | Defectiveness is presumed where the defendant fails to disclose relevant evidence under Article 9(1); where the claimant demonstrates that the product does not comply with mandatory product safety requirements laid down in Union or national law intended to protect against the risk of the damage suffered; or where the claimant demonstrates that the damage was caused by an obvious malfunction during reasonably foreseeable use or under ordinary circumstances. |
| 10(4) and (5); recital 48 | A court shall presume defectiveness or causation, or both, where, despite disclosure and taking into account all the relevant circumstances, the claimant faces excessive difficulties, in particular due to technical or scientific complexity, and demonstrates that defectiveness or causation is likely. Recital 48 names machine learning as complex technology and says a claimant concerning an AI system should not be required to explain the system’s specific characteristics for the court to find excessive difficulties. The defendant may rebut every presumption. |
| 11(1)(c) and (e), 11(2), 18 | The later-defect defense does not apply where the defect is due to a related service, software including updates or upgrades, a lack of software updates or upgrades necessary to maintain safety, or a substantial modification, provided that it is within the manufacturer’s control. The state-of-the-art defense remains, subject to Member State derogations under Article 18. |
| 12, 15 | Two or more operators liable for the same damage can be held liable jointly and severally. Liability may not, in relation to the injured person, be limited or excluded by a contractual provision or by national law. |
| 16, 17 | Three-year limitation period from when the injured person became aware, or should reasonably have become aware, of the damage, the defectiveness and the identity of the liable operator. Claims expire 10 years after the product was placed on the market or put into service (or, if substantially modified, made available or put into service again), or 25 years where the latency of a personal injury prevented earlier proceedings. |

## What happened to the EU AI Liability Directive, and how does the AI Act fit?

The Commission’s 2022 proposal on adapting non-contractual civil liability rules to artificial intelligence, COM(2022) 496, is withdrawn. Annex IV of the Commission work programme 2025 (February 11, 2025) listed it with the reason “No foreseeable agreement”. The Commission approved the withdrawal on July 16, 2025, and the notice was published in the Official Journal on October 6, 2025 (C/2025/5423).

The AI Act sets duties and administrative fines. The individual remedies it provides are a complaint and an explanation, not damages:

- Anyone with grounds to consider that the Act has been infringed may complain to the relevant market surveillance authority, without prejudice to other administrative or judicial remedies (Article 85).
- An affected person subject to a deployer’s decision based on the output of an Annex III high-risk system, other than point 2, that produces legal effects or similarly significantly affects them in a way they consider adverse to their health, safety or fundamental rights may obtain clear and meaningful explanations of the role of the system in the decision-making procedure and the main elements of the decision, subject to exceptions or restrictions that follow from Union or national law in compliance with Union law, and only to the extent Union law does not otherwise provide that right (Article 86). Article 86 has applied since August 2, 2026 but reaches a high-risk system only once its high-risk duties apply.
The Act does not affect the liability rules for intermediary services in Chapter II of the Digital Services Act (Article 2(5)). Since July 27, 2026, Article 110 has listed the Act in Annex I of the Representative Actions Directive, (EU) 2020/1828, so qualified entities can seek redress, such as compensation as available under Union or national law, for traders’ infringements that harm or may harm the collective interests of consumers. In a product liability claim, non-compliance with mandatory product safety requirements triggers the Article 10(2)(b) presumption in the table above; the Product Liability Directive does not say which AI Act requirements count, so that will be for the courts. The [EU AI Act compliance guide](https://superintelligencecompliance.ai/guides/eu-ai-act-compliance) covers the duties.

## Can an AI developer face product liability in the US?

An AI developer can be sued on that theory, and the claims can survive a motion to dismiss. In *Garcia v. Character Technologies, Inc.*, No. 6:24-cv-01903 (M.D. Fla.), a mother alleged that the Character A.I. app caused her son’s death and sued its developer, its founders and Google. In an order filed May 21, 2025, the court dismissed the intentional infliction of emotional distress claim without leave to amend and otherwise denied the motions to dismiss. On the allegations taken as true, it:

- said it was not prepared to hold, at that stage, that the app’s output is speech, so it did not dismiss the claims on First Amendment grounds;
- held the app is a product for the purposes of the product liability claims so far as those claims arise from defects in the app rather than ideas or expressions within it;
- found the complaint sufficiently alleged that Google was liable as a component part manufacturer and for aiding and abetting.
The parties reported a mediated settlement in principle on January 7, 2026, and the court dismissed the case with prejudice on October 19, 2026, so no appellate court reviewed these rulings.

No federal statute makes AI a product. The AI LEAD Act (S. 2937), introduced September 29, 2025, finds that AI systems are products and would create a federal cause of action against developers and deployers. It has not passed; as of November 15, 2026, its latest action is referral to the Senate Judiciary Committee.

## Can an AI vendor be liable for its customer’s decisions?

Possibly, as its customer’s agent. That theory has survived a motion to dismiss, but no court on this page has found a vendor liable on it. In *Mobley v. Workday*, No. 3:23-cv-00770 (N.D. Cal.), the court’s July 12, 2024 order let the federal Title VII, ADEA and ADA claims proceed because the complaint plausibly alleged Workday’s liability as an agent of employers; it dismissed the employment agency theory and the intentional discrimination claims. On June 22, 2026, ruling on a later motion to dismiss, it rejected Workday’s argument that its FEHA liability as an agent turns on its customers’ liability: under the California Supreme Court’s *Raines* decision, an agent’s liability is direct, for its own engagement in FEHA-regulated activities on the employer’s behalf, not derivative.

On May 16, 2025, the court preliminarily certified a nationwide collective of applicants age 40 and over. The case was still pending as of November 2026, and nothing had been adjudicated against Workday; the class certification hearing is set for March 9, 2027. Using a vendor does not shield the employer: California’s Civil Rights Council regulations define an employer’s agent to include anyone acting on the employer’s behalf, directly or indirectly, including through an automated-decision system. See the [employment guide](https://superintelligencecompliance.ai/guides/ai-compliance-employment-hr).

## Can a business blame its chatbot or its AI?

Not in the decisions and statutes reviewed here. In *Moffatt v. Air Canada*, 2024 BCCRT 149 (February 14, 2024), a small claims decision of British Columbia’s Civil Resolution Tribunal, a customer relied on the airline’s website chatbot, which suggested the customer could apply for bereavement fares retroactively. The tribunal rejected the argument that the airline could not be liable for information from its chatbot, found it responsible for all the information on its website whether from a static page or a chatbot, and ordered it to pay $812.02, of which $650.88 was damages.

California and Utah have put the point in statute, and the table above quotes both rules. Utah’s reaches only violations of statutes the Division of Consumer Protection administers and enforces; SB 226 enacted it as Utah Code 13-75-102, and as of November 15, 2026 the Legislature’s code site shows it as 13-77-102, amended by Chapter 95 of the 2026 General Session.

## Does Section 230 protect AI-generated content?

No decision cited here settles that. Section 230(c)(1) bars treating a provider or user of an interactive computer service as the publisher or speaker of information provided by another information content provider, defined as any person or entity responsible, in whole or in part, for the creation or development of information (47 U.S.C. 230(f)(3)). The question is whether a developer is an information content provider of its own model’s output. In *Bouck v. Meta Platforms* (N.D. Cal.), a district court held at the pleading stage on March 24, 2026 that Section 230 did not require dismissing claims alleging that Meta’s generative AI advertising tools wrote fraudulent ad language; on June 11, 2026 it dismissed the case without prejudice on a securities-law ground. The *Garcia* order did not address Section 230.

The other Section 230 ruling cited here also concerns platforms. In *Kohls v. Bonta*, No. 2:24-cv-02527 (E.D. Cal.), a final judgment on August 20, 2025 declared that AB 2655 violates and is preempted by Section 230 as applied to X Corp. and Rumble and permanently enjoined its enforcement against them. The state appealed to the Ninth Circuit on September 19, 2025 (No. 25-6138); we found no appellate decision as of November 16, 2026. See the [deepfake laws guide](https://superintelligencecompliance.ai/guides/deepfake-laws).

## Which state AI laws create defenses or shift liability?

Of the state AI laws this site covers, these do, as does Utah’s rule above; the table quotes each:

- **Texas** has the Sec. 552.105(e) defense quoted in the table, but only the Attorney General enforces Chapter 552, apart from licensing-agency sanctions, and there is no private right of action. The framework clause attaches to the internal review route. See the [Texas guide](https://superintelligencecompliance.ai/guides/texas-ai-law-hb-149).
- **Colorado** has no framework-based defense. Its relative-fault and void-indemnity rules are quoted in the table. A federal court order in *X.AI LLC v. Weiser* bars the Attorney General from initiating enforcement, including an investigation, for violations occurring on or before 14 days after the court rules on xAI’s preliminary injunction motion. As of November 15, 2026, that motion had not been filed and the implementing rules had not been adopted. See the [Colorado guide](https://superintelligencecompliance.ai/guides/colorado-sb-26-189).
- **California** bars the autonomy defense (AB 316) and gives a person injured in fact by a violation a private right of action against companion chatbot operators (SB 243). See the [California guide](https://superintelligencecompliance.ai/guides/california-ai-laws).

## What liability do lawyers face when they use AI?

Professional discipline and court sanctions, under the existing rules. ABA Formal Opinion 512 (July 29, 2024) adds no rule; it says lawyers must have a reasonable understanding of the capabilities and limitations of the tool, that relying on output without an appropriate degree of review could violate the duty of competence, and that before filing, lawyers must review AI output, including citations, and correct errors. It makes the lawyer, not the vendor, answerable for confidentiality and supervision.

Courts sanction the lawyer who signs for fake citations, whatever tool drafted the brief: $5,000 jointly and severally on two lawyers and their firm in *Mata v. Avianca* (S.D.N.Y., June 22, 2023), and $15,000 from each of two lawyers, plus the appellees’ attorney fees and double costs, and referral for possible discipline in *Whiting v. City of Athens* (6th Cir., March 13, 2026). See the [law firms page](https://superintelligencecompliance.ai/industries/law-firms-legal).

## How should contracts allocate AI liability?

Under the Product Liability Directive, contracts move cost between companies but do not cut what an injured person can recover. Liability may not, in relation to the injured person, be limited or excluded by contract (Article 15), and recourse between operators follows national law (Article 14). Article 12(2) deals with recourse between operators: a manufacturer that integrates a software component has no right of recourse against the manufacturer of a defective software component that causes damage if that manufacturer was a microenterprise or small enterprise (counting partner and linked enterprises) when it placed the component on the market and the two contractually agreed to waive that right. In Colorado, a contract clause that indemnifies a developer or deployer for its own violations of Colorado anti-discrimination law is void (6-1-1707). The [AI vendor risk assessment guide](https://superintelligencecompliance.ai/guides/ai-vendor-risk-assessment) covers the clauses.

## AI liability checklist

- List each product with software or AI that will be placed on the EU market or put into service on or after December 9, 2026, and name who would answer for it under Article 8
- Keep design, testing and update records you could produce under an Article 9 disclosure order, with a plan to protect trade secrets
- Plan safety updates for the product’s life; where a defect is due to a lack of updates necessary to maintain safety that is within your control, the later-defect defense does not apply (Article 11(2))
- Drop “the AI acted on its own” from defense playbooks in California actions and in matters under statutes Utah’s Division of Consumer Protection administers and enforces
- Treat chatbot answers as your own statements and review what customer-facing bots say about prices, refunds and policies
- For screening tools, assume both vendor and customer can be sued, and test for adverse impact before use
- Document which Texas Sec. 552.105(e) route you would rely on in an Attorney General action and, for the internal review route, the framework you substantially comply with; remove indemnities Colorado 6-1-1707 voids
- Require lawyers and other professionals to verify AI output, including every citation, before it is filed or relied on
- Match vendor contracts to what the law lets you shift, and confirm insurance for what it does not
How this fits the other guides: the [AI compliance guide](https://superintelligencecompliance.ai/guides/ai-compliance-guide) is the hub for how existing law reaches AI, and the [manufacturers page](https://superintelligencecompliance.ai/industries/manufacturers) sets the Product Liability Directive beside the Machinery Regulation and the Cyber Resilience Act.

This guide is general information as of November 15, 2026, not legal advice. Confirm the current law with qualified counsel, the court docket or the regulator before acting.

## Sources

- [Directive (EU) 2024/2853 on liability for defective products (Articles 2 to 4, 6 to 18, 21, 22; recitals 13, 48)](https://eur-lex.europa.eu/eli/dir/2024/2853/oj), Official Journal of the European Union, November 18, 2024.
- [Corrigendum to Directive (EU) 2024/2853 (Article 2(1)), OJ L, 2026/90364](https://eur-lex.europa.eu/eli/dir/2024/2853/corrigendum/2026-05-07/oj), Official Journal of the European Union, May 7, 2026.
- [Withdrawal of Commission proposals (C/2025/5423)](https://eur-lex.europa.eu/eli/C/2025/5423/oj), Official Journal of the European Union, October 6, 2025.
- [Procedure 2022/0303(COD), COM(2022) 496, AI Liability Directive: proposal withdrawn](https://eur-lex.europa.eu/procedure/EN/2022_303), EUR-Lex, accessed November 15, 2026.
- [Annexes to the Commission work programme 2025, COM(2025) 45 final, Annex IV (PDF)](https://commission.europa.eu/document/download/7617998c-86e6-4a74-b33c-249e8a7938cd_en?filename=COM_2025_45_1_annexes_EN.pdf), European Commission, February 11, 2025.
- [Regulation (EU) 2024/1689 (Artificial Intelligence Act), consolidated text as of July 27, 2026 (Articles 2(5), 85, 86, 110, 113)](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:02024R1689-20260727), EUR-Lex, accessed November 15, 2026.
- [Directive (EU) 2020/1828 on representative actions for the protection of the collective interests of consumers (Articles 2, 9; Annex I)](https://eur-lex.europa.eu/eli/dir/2020/1828/oj), Official Journal of the European Union, December 4, 2020.
- [*Garcia v. Character Technologies, Inc.*, No. 6:24-cv-01903, Order on motions to dismiss, ECF No. 115 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.flmd.433581/gov.uscourts.flmd.433581.115.0.pdf), U.S. District Court for the Middle District of Florida, May 21, 2025, via CourtListener RECAP.
- [*Garcia v. Character Technologies, Inc.*, Notice of Resolution, ECF No. 242 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.flmd.433581/gov.uscourts.flmd.433581.242.0.pdf), January 7, 2026, via CourtListener RECAP.
- [*Garcia v. Character Technologies, Inc.*, No. 6:24-cv-01903 (M.D. Fla.), docket, ECF Nos. 278 and 280](https://www.courtlistener.com/docket/69300919/garcia-v-character-technologies-inc/), CourtListener, accessed November 15, 2026.
- [S. 2937, AI LEAD Act](https://www.congress.gov/bill/119th-congress/senate-bill/2937), Congress.gov, accessed October 21, 2026.
- [Bill status, S. 2937 (119th Congress), XML](https://www.govinfo.gov/bulkdata/BILLSTATUS/119/s/BILLSTATUS-119s2937.xml), GovInfo bulk data, accessed November 15, 2026.
- [*Mobley v. Workday, Inc.*, No. 3:23-cv-00770, Order granting in part and denying in part motion to dismiss, ECF No. 80 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.cand.408645/gov.uscourts.cand.408645.80.0.pdf), U.S. District Court for the Northern District of California, July 12, 2024, via CourtListener RECAP.
- [*Mobley v. Workday, Inc.*, Order granting in part and denying in part motion to dismiss, ECF No. 360 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.cand.408645/gov.uscourts.cand.408645.360.0.pdf), June 22, 2026, via CourtListener RECAP.
- [*Mobley v. Workday, Inc.*, No. 3:23-cv-00770 (N.D. Cal.), docket, ECF Nos. 418, 449 and 450](https://www.courtlistener.com/docket/66831340/mobley-v-workday-inc/), CourtListener, accessed November 15, 2026.
- [Civil Rights Council Secures Approval for Regulations to Protect Against Employment Discrimination Related to Artificial Intelligence](https://calcivilrights.ca.gov/2025/06/30/civil-rights-council-secures-approval-for-regulations-to-protect-against-employment-discrimination-related-to-artificial-intelligence/), California Civil Rights Department, June 30, 2025.
- [*Moffatt v. Air Canada*, 2024 BCCRT 149 (PDF)](https://decisions.civilresolutionbc.ca/crt/crtd/en/525448/1/document.do), Civil Resolution Tribunal of British Columbia, February 14, 2024.
- [AB-316 Artificial intelligence: defenses (Chapter 672, Statutes of 2025)](https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260AB316), California Legislative Information, 2025 (chaptered text read from the Legislative Counsel bulk file, pubinfo_2025).
- [S.B. 226 Artificial Intelligence Consumer Protection Amendments (enrolled)](https://le.utah.gov/Session/2025/bills/enrolled/SB0226.pdf), Utah Legislature, 2025.
- [Utah Code Title 13, Chapter 77, Generative Artificial Intelligence: Consumer Disclosures and Enforcement (PDF)](https://le.utah.gov/xcode/Title13/Chapter77/C13-77_2025050820250508.pdf), Utah Legislature, accessed November 15, 2026.
- [H.B. No. 149 (enrolled), Texas Responsible Artificial Intelligence Governance Act](https://capitol.texas.gov/tlodocs/89R/billtext/html/HB00149F.htm), Texas Legislature, 2025.
- [SB 26-189, Session Laws 2026, Chapter 131 (enrolled act)](https://leg.colorado.gov/laws/session-laws/SB26-189/131/download), Colorado General Assembly, approved May 14, 2026.
- [*X.AI LLC v. Weiser*, No. 1:26-cv-01515 (D. Colo.), docket, minute order ECF No. 24 (April 27, 2026) and ECF No. 25 (October 2, 2026)](https://www.courtlistener.com/docket/73171074/x-ai-llc-v-weiser/), CourtListener RECAP, accessed November 15, 2026.
- [SB-243 Companion chatbots (Chapter 677, Statutes of 2025)](https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB243), California Legislative Information, 2025.
- [47 U.S.C. 230, Protection for private blocking and screening of offensive material](https://www.govinfo.gov/content/pkg/USCODE-2023-title47/html/USCODE-2023-title47-chap5-subchapII-partI-sec230.htm), U.S. Code (2023 edition), GovInfo.
- [*Kohls v. Bonta*, No. 2:24-cv-02527, Order and Final Judgment and Permanent Injunction as to AB 2655, ECF No. 98 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.caed.453046/gov.uscourts.caed.453046.98.0.pdf), U.S. District Court for the Eastern District of California, August 20, 2025.
- [*Kohls v. Bonta*, Notice of Appeal, ECF No. 106 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.caed.453046/gov.uscourts.caed.453046.106.0.pdf), U.S. District Court for the Eastern District of California, September 19, 2025.
- [*The Babylon Bee, LLC v. Bonta*, No. 25-6138, docket (latest entry No. 77, August 12, 2026; no opinion)](https://www.courtlistener.com/docket/71503636/the-babylon-bee-llc-et-al-v-bonta-et-al/), U.S. Court of Appeals for the Ninth Circuit, via CourtListener RECAP, accessed November 16, 2026.
- [*Bouck v. Meta Platforms, Inc.*, No. 3:25-cv-05194-RS, Order denying in part and granting in part motion to dismiss second amended complaint with leave to amend, ECF No. 61 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.cand.451567/gov.uscourts.cand.451567.61.0.pdf), U.S. District Court for the Northern District of California, March 24, 2026, via CourtListener RECAP.
- [*Bouck v. Meta Platforms, Inc.*, Order granting motion to dismiss without prejudice, ECF No. 83 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.cand.451567/gov.uscourts.cand.451567.83.0.pdf), June 11, 2026, via CourtListener RECAP.
- [ABA Formal Opinion 512, Generative Artificial Intelligence Tools (PDF)](https://www.mcle.org/ePrograms/2250081P01/1--aba-formal-opinion-512.pdf), American Bar Association Standing Committee on Ethics and Professional Responsibility, July 29, 2024 (copy hosted by MCLE).
- [Mata v. Avianca, Inc., No. 1:22-cv-01461 (S.D.N.Y.), Opinion and Order on Sanctions, ECF No. 54 (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.nysd.575368/gov.uscourts.nysd.575368.54.0.pdf), June 22, 2023, via CourtListener RECAP.
- [Whiting v. City of Athens, Nos. 24-5918/5919, 25-5424 (6th Cir.) (PDF)](https://storage.courtlistener.com/recap/gov.uscourts.ca6.152857/gov.uscourts.ca6.152857.50.2.pdf), March 13, 2026, via CourtListener RECAP.

## Common questions

**Is AI a product under EU product liability law?**

Yes, for products placed on the market or put into service on or after December 9, 2026 (Article 2(1), as corrected on May 7, 2026). Directive (EU) 2024/2853 defines product to include software, and recital 13 says AI system providers within the meaning of the AI Act should be treated as manufacturers. Free and open-source software developed or supplied outside the course of a commercial activity is excluded.

**Can a company argue that its AI acted on its own?**

Not in California, where Civil Code 1714.46 (AB 316), in force since January 1, 2026, bars a defendant who developed, modified, or used AI alleged to have caused a harm to the plaintiff from asserting that the AI autonomously caused the harm to the plaintiff; the defendant may still present any other affirmative defense, including evidence relevant to causation or foreseeability, and other evidence relevant to the comparative fault of any other person or entity. Utah provides that it is not a defense to a violation of a statute its Division of Consumer Protection administers and enforces that generative AI made the violative statement, undertook the violative act, or was used in furtherance of the violation.

**What happened to the EU AI Liability Directive?**

It was withdrawn. The Commission listed the proposal, COM(2022) 496, for withdrawal in its 2025 work programme, citing no foreseeable agreement, approved the withdrawal on July 16, 2025, and published the notice in the Official Journal on October 6, 2025 (C/2025/5423). For products placed on the market or put into service on or after December 9, 2026, compensation claims for defective AI products run through the revised Product Liability Directive as transposed into national law; Directive 85/374/EEC continues to apply to products placed on the market or put into service before that date, and fault-based claims under national law remain available.

