The primary statute governing product liability in the UK is the Consumer Protection Act 1987 ('the Act'), which implements strict liability for defective products causing personal injury or property damage. Drafted over 40 years ago, it pre-dates the digital economy, artificial intelligence, and connected devices entirely, and is widely considered ill-suited to address the risks posed by modern technology.

Following Brexit, the UK now also faces growing divergence from the EU, which updated its Product Liability Directive in 2024 to explicitly cover software and AI. It is against this backdrop that the Law Commission is reviewing this framework, with its consultation period running until 14 January 2027. 

Key drivers of reform

The recognised limitations with the Act as it currently stands include: 

• AI and software: Whether software constitutes a 'product' under the Act remains unresolved, creating fundamental uncertainty about the Act's application to some of the most consequential technologies in use today.
• The development risks defence: Producers can escape liability where the defect could not have been discovered given the state of knowledge at the time of supply. Critics argue this is unduly generous, particularly for AI systems that evolve post-deployment.
• Post-sale updates: Where a product becomes defective following a remote software update, the Act provides no clear answer on liability.
• Burden of proof: Claimants must prove defect, damage, and causation. This is a high evidential bar for technically complex AI products.
• Compensation thresholds: The £275 damage threshold currently contained within the Act may operate unfairly in the context of digital product harms such as data loss.

The Law Commission's key proposals

The current proposals would incorporate: 

• Extending 'product' to cover software and AI: Including systems delivered via the cloud or as a service, whether embedded in hardware or standalone.  
• Reforming the defect standard: The standard incorporating the context of use, foreseeable misuse, reasonable safety expectations, and stated purpose of the digital product. 
• Post-market update liability: Liability to attach to substantial modifications made after initial market placement that introduce new risks. 
• Restricting the development risks defence: Particularly for AI, given that failures are often foreseeable even if not specifically anticipated.
• Supply chain and multi-party liability: Developers, deployers, and those who substantially modify AI systems to bear appropriate liability, not solely the original manufacturer.
• Review of compensation thresholds: To ensure the £275 threshold and other limitations do not operate unfairly in digital harm contexts.

How would insurers be affected by the proposed reforms?

It seems inevitable that the legislative framework that determines liability when a product causes harm will shortly undergo significant change, and insurers will want to understand and respond to the implications across a number of areas.

• Broader exposure and rising claims frequency: Extending strict liability to software and AI - combined with a relaxed development risks defence and a lowered evidential burden for claimants - will almost certainly increase the volume and value of claims. Products that previously sat outside the scope of the Act, or whose defects were previously challenging for claimants to prove, will increasingly give rise to successful claims. Insurers should expect to see not only more claims, but claims of greater complexity and cost.
• Policy gaps and coverage adequacy: Much of the existing product liability market was constructed around physical goods. The extension of liability to software, AI systems, and cloud-delivered services exposes potential gaps in policy wordings that were never designed to respond to these risks. Particularly challenging areas include AI-driven decision-making that causes harm, and post-sale software updates that render a product defective - both of which may fall outside the coverage contemplated by existing policy language.
• Challenge of pricing dynamic and evolving risks: AI products present underwriters with a genuinely novel challenge. Unlike a static physical product, an AI system's risk profile can change materially after deployment - through updates, retraining, or changes in the environment in which it operates. Add to this the potential for systemic, large-scale losses where a single defective system is used at scale, and the difficulty of accurately pricing and reserving becomes considerable. The result is likely to be upward pressure on premiums and, in some segments of the market, reduced appetite for cover.
• Blurring of policy boundaries: The proposed reforms do not operate in a vacuum. As product liability rules expand to capture AI and software-related harms, the boundaries between product liability, cyber, professional indemnity, and directors' and officers' insurance may blur. Disputes about which policy responds - and to what extent - are likely to become more frequent, placing greater demands on coverage analysis and potentially leading to coverage disputes between insurers.
• Systemic and accumulation risk: Where a defective AI system is deployed across a large user base, the resulting claims could be vast in both number and aggregate value. This presents a risk beyond the capacity of individual insurers and could have significant implications across the wider insurance market. 

What should insurers do?

A reactive approach to the proposed reforms carries real risk. Insurers will want to take practical steps now, including:

• Auditing existing policy wordings for gaps in AI and software cover. 
• Developing products built specifically for technology risks. 
• Investing in the technical underwriting expertise needed to assess those risks properly. 
• Engaging with reinsurers early on capacity and structure given the potential for large-scale, systemic losses.

Conclusion

The proposed reforms represent one of the most significant shifts in UK product liability law in a generation. The direction of travel is clear - broader liability, and losses that are larger and more complex than the market has historically seen. Insurers who act ahead of the legislative curve will be far better placed to manage the implication when the law is settled.