The United Kingdom has decided not to rush a new copyright framework for artificial-intelligence training. In a report published on 18 March 2026, the government said it would not introduce reforms until it was confident that they could protect right holders, support access to high-quality content for AI developers and deliver wider economic benefits.
That is a meaningful change in direction. A broad text-and-data-mining exception with an opt-out had been the government’s preferred consultation proposal. The 2026 report says that approach is no longer preferred after strong opposition, limited evidence and continued uncertainty about licensing markets, litigation and international rules.
What the consultation revealed
The consultation ran from December 2024 to February 2025 and received 11,520 responses. Participants included creators, performers, publishers, AI developers, researchers, cultural organisations, small businesses and legal professionals.
Four broad options were considered:
- keep the current law;
- require licensing in all cases;
- introduce a broad data-mining exception;
- introduce an exception with an opt-out and transparency measures.
The government found no consensus. Many right holders argued that an opt-out would be difficult to use and could weaken the value of their work. Many technology-sector respondents argued that wider exceptions could make the UK more competitive. The report acknowledges that both creative production and AI development matter, while also admitting that the evidence needed to balance them remains incomplete.
No immediate change to copyright law
The report is a policy paper and impact assessment, not legislation. It does not create a new exception and it does not remove existing rights. The government’s stated position is to gather more evidence, examine alternative policy approaches and monitor technology, court cases, licensing markets and international developments.
For creators, the practical lesson is simple: do not treat headlines about a consultation as if the law has already changed. For developers, the absence of a new reform is not permission to ignore copyright. The applicable law, licences and terms governing source material still matter.
Transparency moves to the centre
The report says greater transparency about the content and data used to train AI systems can help right holders assert their rights. It proposes working with industry and experts on best practice for input transparency while monitoring approaches adopted elsewhere.
Transparency is not a solved technical problem. Training datasets can contain billions of items, supply chains are complex and different stakeholders disagree about the right level of disclosure. Still, the direction is clear: provenance, documentation and auditable processes are becoming central to the relationship between AI systems and creative work.
Licensing remains market-led—for now
The government describes licensing as important both for compensating creators and for giving AI developers legitimate access to useful material. It does not propose intervening directly in the licensing market at this stage. Instead, it plans to watch how market-led models evolve and to examine tools such as the Creative Content Exchange.
That leaves a difficult gap for independent creators and small businesses. Large organisations can negotiate complex agreements; individual artists may have less bargaining power and less visibility into how their work is used. The report recognises concerns about fair outcomes but does not yet provide a final mechanism.
Labels, technical tools and digital replicas
The report also considers labelling AI-generated material, technical signals used by right holders, enforcement and realistic replicas of a person’s face or voice. It proposes further work rather than immediate general rules. It also distinguishes wholly computer-generated work from human work created with AI assistance—a distinction likely to matter more as creative workflows become hybrid.
What responsible teams can do now
- Keep a record of the sources, licences and permissions behind training or reference material.
- Separate human-created, AI-assisted and fully generated assets in internal production records.
- Do not assume that publicly accessible content is automatically free to use for every purpose.
- Review vendor claims about training data and indemnity instead of relying on marketing language.
- Preserve the ability to update provenance records as law and platform rules evolve.
The Mythic Mode perspective
Creative technology works best when experimentation and authorship are both visible. A durable digital culture needs room for new tools, but it also needs evidence: who made an asset, what material informed it and what permissions apply. The UK report does not settle those questions. It makes clear why shortcuts are unlikely to produce a trustworthy answer.
This article provides general information, not legal advice. Copyright rules depend on jurisdiction and facts; check current law and professional advice before making legal or licensing decisions.