The UK Jurisdiction Taskforce (UKJT) published its Legal Statement on Liability for AI Harms on 7 July 2026, following a public consultation conducted from January 2026. The Statement constitutes the UKJT's authoritative legal analysis of how the private law of England and Wales allocates liability for non-deliberate harm arising from AI systems. It carries persuasive authority in English courts addressing novel AI disputes and operates alongside, rather than displacing, sector-specific regulatory liability regimes.
The Statement applies established English tort principles — primarily negligence, product liability, and the law of misrepresentation — to AI harm scenarios. Professionals are subject to a negligence standard that encompasses both over-reliance on AI output and failure to adopt AI where a competent practitioner in their field would have done so. Application developers who deploy AI in a specific, known context owe a duty of care where harm is foreseeable; that duty narrows progressively up the AI supply chain. Foundation model developers are, in most circumstances, unlikely to bear liability for downstream harms caused by unforeseeable uses of general-purpose models. Where a business deploys a chatbot that produces false information, the deploying business cannot disclaim liability on the ground that the AI system rather than a human employee generated the inaccurate output.
Legal practitioners, AI system deployers, enterprises integrating AI into client-facing workflows, and professional services firms face the most direct implications. Lawyers, doctors, accountants, and financial advisers using AI tools in client matters are held to a professional negligence standard that requires critical evaluation of AI output and, in appropriate circumstances, active use of available AI tools. Companies deploying application-layer AI bear residual liability to third-party claimants for foreseeable harms even where the underlying model is third-party supplied; contractual indemnities between application developers and foundation model providers do not affect claims by injured third parties.
The Statement leaves open the liability treatment of AI agents operating autonomously across multi-step tasks, the attribution of harm where multiple AI systems in a supply chain each contribute to damage, and the handling of AI-generated professional output where no human decision point intervenes in the chain. These are flagged for future UKJT analysis or legislative consideration. The Statement does not address criminal liability, regulatory liability under sector-specific AI regimes, or liability governed by the Consumer Rights Act 2015.
Licentium may advise technology companies, professional services firms, and AI deployers on AI liability risk analysis and has a partner network to assist clients in England and Wales and comparable jurisdictions. Organisations reviewing AI deployment frameworks following the UKJT statement are welcome to contact us. Work we undertake includes AI liability risk assessment, professional negligence exposure analysis, AI governance documentation, supplier contract structuring, and cross-border AI compliance review.
Source: UK Jurisdiction Taskforce, Legal Statement on Liability for AI Harms, 7 July 2026