When an AI system causes injury or damage, the injured person usually wants to sue the company behind it. Whether they need to prove carelessness, or only that the product was defective, depends on where they are.
European Union: strict liability for software
The revised Product Liability Directive (EU) 2024/2853 applies to products placed on the EU market from 9 December 2026. Software, including AI systems, is expressly a product. Liability is strict, so there is no need to prove fault, only a defect, damage and a causal link. The directive makes claims easier for technically complex products and can reach importers, authorised representatives and, in some cases, online platforms. Damage to data and psychological health can count. A product can become defective through a missing update or a cybersecurity weakness.
United Kingdom
The Consumer Protection Act 1987 still governs. Whether software without a physical carrier is a “product” under it is unsettled, so claimants often rely on negligence or contract instead. A UK supplier selling into the EU will face the new directive there.
United States
Product liability is state law. Courts are deciding whether AI applications are products or services. In a Florida case against a companion chatbot company, a federal judge in 2025 allowed product liability claims to proceed; the case later settled.
In the EU, “we only supply software” stops being a defence in December.
What to do
If you place software or AI-enabled products on the EU market, review your update policy, technical documentation and supplier indemnities now. Check whether your liability insurance covers strict liability for software.
This article is general information on the law as at November 2026. It is not legal advice and has not been updated since that date. See the terms of use.