UK shelves AI copyright reform for now
Creators and publishers get no new rules on AI training, and contracts carry the weight.
Background
On 18 March 2026 the UK government published its long-awaited report on copyright and AI. It concluded that reform is not being introduced at this point, having consulted extensively and received significant opposition from the creative sector.
Three key points
1. No preferred option. The government confirmed it no longer has a preferred approach to the tension between copyright protection and AI training. It dropped its earlier preference for a broad text and data mining exception with an opt-out. No new regulator for AI and copyright is proposed.
2. Proposed changes to computer-generated works. The report proposes ending specific copyright protection for works that are wholly computer-generated, while retaining protection for works made with AI assistance by a human author. This distinction matters for content studios, design agencies and publishers building AI into their production processes.
3. The High Court ruled on Getty. In the first major UK ruling on generative AI and copyright, the High Court rejected Getty's central claim that Stable Diffusion's model weights constituted an infringing copy of its images, but found limited trademark infringement where AI outputs reproduced visible watermarks.
Why this matters
The absence of reform is itself a position. Contracts and licensing arrangements now carry the full weight of copyright protection for training data, without the clarity that statutory rules would provide. Publishers and rights holders should review how their terms deal with AI system providers.
Sources: Osborne Clarke, March outlook; Regulations.AI.
