The Law Commission of England and Wales has proposed widening product liability law so that software and AI systems, whether delivered on a physical medium or supplied digitally, count as products. Its consultation opened on 8 October 2026 and runs until 14 January 2027.
The proposals would also cover the design components and features of social media and other online platforms, according to the Commission’s summary of its consultation paper. It expects to publish a final report with recommendations to Parliament in late 2027.
Business Pill · POST-MARKET SURVEILLANCE
A short explainer of post-market surveillance: a test shows one day, so safety has to be watched after release. It teaches the general idea only and says nothing about any body or company in this story.
The key insight: As we read it, the proposals tie liability to the moment a product changes, not only the moment it is sold. Producers could not rely on the time-of-supply defence for defects they cause through later updates or missing updates, and the long-stop clock would reset after a substantial modification.
The current regime, Part 1 of the Consumer Protection Act 1987, was designed primarily with tangible goods in mind and does not easily accommodate software and AI systems, the Commission says. It proposes expanding the definition of a producer to include the person who developed them.
The summary lists features associated with AI systems, including machine operation, the generation of outputs from inputs, autonomy in producing those outputs and the possibility of self-learning after deployment, and asks consultees whether they are a helpful basis for the law.
The Commission says it does not consider an AI system’s ordinary self-learning to be a substantial modification of the product.
The Law Commission of England and Wales lists ten key provisional proposals for product liability reform (our count of its list), including treating software and AI systems as products. Consultation runs from 8 October 2026 to 14 January 2027.
Updates, Open Source and Data
Under the proposals, a producer could not rely on the defence that a defect did not exist at the time of supply if it caused the product to become defective afterwards, including by failing to provide necessary software updates.
The Commission does not propose a specific exemption for non-commercial open-source software, but says developers who make it available on a genuinely non-commercial basis would generally fall outside the regime. Open-source software supplied in the course of business would be in scope even where users do not pay for it directly.
It also proposes that destruction and corruption of data, as distinct from data leaks, become compensable damage unless the data is used exclusively for professional purposes.
The long-stop for product liability claims: 10 years from supply under current law, and a proposed 25 years in cases of latent harm, per the Law Commission’s summary of 8 October 2026.
Proving Harm and the Defences
For cases with exceptional evidential difficulties, the Commission proposes a rebuttable presumption of defect. If a defendant chooses to accept the alleged defect rather than rebut it, the starting point would be an award of 50% of the damages that would otherwise have been recoverable.
The development risks defence would be retained for all products and industries, with two updates, the summary says. The 10-year long-stop period would reset after a substantial modification and be extended to 25 years for latent harm.
The Structural Read
Developers become producers. The Commission proposes expanding the definition of a producer to include the person who developed software, AI systems and platform features.
Free is not the test. Open-source software supplied in the course of business would be in scope even where users do not pay for it directly, the summary says.
Hard cases get a new route. Where evidence is exceptionally hard to obtain, a rebuttable presumption of defect could move the burden to the defendant.
Law Commission of England and Wales, 8 October 2026
“We do not consider an AI systemβs ordinary self-learning would satisfy these criteria.”
Three Implications
CONSULTATION TO 14 JANUARY Responses are invited until 14 January 2027, with a final report expected in late 2027.
UPDATES CARRY LIABILITY Under the proposals, a producer that caused a defect after supply, including by failing to provide necessary software updates, could not rely on the time-of-supply defence.
DATA LOSS COUNTS Destruction and corruption of data would become compensable unless the data is used exclusively for professional purposes.
The framework’s starting point: “The governance layer sits at the intersection of capability, regulation, national security, access control, and geopolitical alignment.”
As we read it, the consultation adds a liability question to that layer: which party answers for harm from an AI system once it has been updated, modified or built into another product.
What Is Not Established
We read the Law Commission’s project page and its summary of the consultation paper; we did not read the full consultation paper. These are provisional proposals for England and Wales, open to consultation, and any change in the law would need the Government and Parliament. We did not contact the Commission.
Business Engineer Framework
Inside Anthropic’s Permission Layer
A Business Engineer framework on the governance layer of the AI stack: who decides access, use and oversight.
The Law Commission has proposed that software and AI systems be treated as products under product liability law in England and Wales, with producers unable to rely on the time-of-supply defence for defects they cause through updates, and a rebuttable presumption of defect in hard-to-prove cases. Its consultation runs until 14 January 2027.
A note on sourcing. We read the Law Commission’s project page and its summary of the consultation paper of 8 October 2026; we did not read the full consultation paper. The count of proposals is our count of the Commission’s list. This is not legal advice. We did not contact the Commission. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro Cuofano is a business analyst with an International MBA in corporate finance and business strategy (2012). He leads research at FourWeekMBA, which he created, and at The Business Engineer (businessengineer.ai), his newsletter on AI and business strategy, with over 90,000 subscribers and more than 660 published analyses. His work breaks down how companies make money, read from their own filings: revenue and profit structure, capital allocation and ownership. FourWeekMBA reached about four million business people, including C-level executives, investors, analysts and product managers, in 2022 alone. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
