The UK Jurisdiction Taskforce (UKJT)—part of the LawtechUK initiative backed by the Ministry of Justice—has issued a landmark Legal Statement on Private Law Liability for Artificial Intelligence Harms. The statement definitively concludes that existing English common law principles of tort, contract, and negligence are robust, flexible, and fully equipped to allocate liability when autonomous AI systems cause harm or economic loss.
English common law meets autonomous AI
Addressing growing debate over whether autonomous AI agents require separate legal personhood or bespoke statutory liability regimes, the UKJT affirmed that no new emergency legislation is necessary:
- No separate legal personality: AI systems are tools, not independent legal entities. Liability rests squarely on the human beings or corporate legal entities that develop, train, deploy, or control them.
- Causation and agency: Established common law doctrines of agency, product liability, and vicarious liability can be applied directly to autonomous decision-making loops and algorithmic execution.
- Contractual bindingness: When an AI agent enters into or performs a contract autonomously on behalf of a principal, the principal remains bound by the resulting legal obligations under standard principles of apparent or actual authority.
Tortious liability and the standard of care
In tort and professional negligence, the UKJT clarified how the traditional “reasonable person” test applies to AI deployment. Businesses cannot evade liability by claiming that an advanced neural network was an unpredictable “black box”:
- Duty of care: Deployers owe a clear duty of care to foresee and mitigate predictable failure modes, hallucinations, and discriminatory outputs.
- Reasonable supervision: Failing to implement adequate human-in-the-loop oversight, testing, or prompt boundaries when deploying generative AI in critical workflows will constitute strong evidence of actionable negligence.
- Product liability: Defective training data or algorithmic instability causing property damage or personal injury falls under established product liability frameworks.
Implications for businesses and in-house counsel
The UKJT statement reinforces London’s position as a premier global forum for resolving AI and commercial technology disputes. For organisations building or adopting AI solutions, key operational takeaways include:
- Robust contractual allocation: Explicitly allocate AI failure risks, indemnity clauses, and performance warranties in vendor and customer contracts.
- Maintain comprehensive audit trails: Document system architecture, safety benchmarks, and deployment parameters to defend against prospective negligence claims.
- Review insurance coverage: Verify that existing professional indemnity and commercial general liability policies adequately cover autonomous algorithmic outputs.