The UK’s position on AI and copyright is no longer an open question. On March 18, 2026, the Department for Science, Innovation and Technology published a report on AI and copyright, laid before Parliament under Section 136 of the Data (Use and Access) Act 2025. The report’s central decision: the previously considered “TDM opt-out” model is out. A Licensing-First framework is in.
This matters because the opt-out model would have allowed AI developers to train on copyrighted material unless rights-holders explicitly excluded their work. The approach Britain is now pursuing flips that logic. According to analysis of the DSIT report confirmed by AI SaaS Writer and corroborated by Birketts’ legal update on the March 2026 report, commercial AI developers are expected to seek licenses for training data. The burden of proof has moved.
The Creative Content Exchange, referenced in the report as a pilot initiative, represents the practical infrastructure for Licensing-First. The CCE is designed to create a market mechanism where rights-holders and AI developers can negotiate licenses. How well that market develops, and how quickly, determines whether Licensing-First functions as a compliance pathway or a compliance obstacle.
The report signals that commercial AI developers are expected to seek licenses for training data, according to analysis of the DSIT document. The specific framing used internally in the report is not confirmed from available source text, what is confirmed is the policy direction, the CCE pilot, and the rejection of the opt-out mechanism across multiple independent sources.
The report is also reported to include plans for a future consultation on digital replicas and a potential personality right, with analysts citing a Summer 2026 timeline. That’s a separate track from the training data question, but it signals that the UK’s AI and IP framework is still being constructed. Summer 2026 is the reported timeline; treat it as indicative, not confirmed schedule.
For compliance and product teams, the immediate question isn’t abstract. Any AI developer training on web-scraped UK content, or deploying models in the UK market, should be assessing their current data provenance. The Licensing-First approach doesn’t come with a statutory enforcement mechanism yet, this is a government report, not enacted legislation. But policy direction at this level of specificity tends to precede formal law, and the Data (Use and Access) Act 2025 has already created the parliamentary mandate for this report’s direction.
Compare this with the US framework’s approach: where Washington signals that training on copyrighted material may not violate copyright law and defers to the courts, London says licensing is the expected path and is building the market infrastructure to support it. These are not compatible postures. A developer operating in both markets faces structurally different signals from two major regulatory environments, and, as the deep-dive below explores, a third from Brussels.
What to watch: the Creative Content Exchange pilot’s operational timeline; any formal legislative steps following the DSIT report; and the Summer 2026 consultation on digital replicas. The GOV.UK official publication of this report is the primary source, see full coverage on the Regulation pillar page. For cross-border analysis, see the deep-dive: Three Governments, Three Copyright Approaches.