On Wednesday, Liz Kendall (the Secretary of State for Science, Innovation and Technology) published a written statement, together with a Report and Impact Assessment, on the Government's 'progress' towards addressing the application of UK copyright law to the training of AI models. 

Since the end of 2024, the Government has "engaged extensively with creatives, AI firms, industry bodies, unions, academics and AI adopters, and that engagement has shaped our approach", has made a strong commitment to "do what is right for the whole British economy" and is "determined to make progress on these issues". 

After all that engagement, what is the Government's long-anticipated approach to resolving the tension between AI developers and copyright owners? Following an extensive consultation process with over 11,500 responses, the establishment of four technical working groups with stakeholders, and a protracted period of lengthy consideration and debate - the Government now "no longer has a preferred option" for legislative reform on AI training. The anticlimactic conclusion: more consultation, for an indefinite period.

Background

The report and impact assessment follow the Copyright and AI consultation, which was launched by the Government in December 2024. The consultation's aim was to resolve the ongoing tension between copyright law and artificial intelligence, and to consider how UK law should balance the interests of copyright owners against the interests of AI developers. In its initial consultation document, the government set out a preferred option: a new text-and-data mining (TDM) exception to copyright that would allow AI companies to use copyright-protected works for AI training purposes, provided the rightsholder had not "reserved their rights" (or "opted out").

The 2024 consultation's preferred option of a broad TDM copyright exception subject to opt-out was intended to bring UK copyright law broadly in line with the EU approach, but was strongly opposed by the UK's creative industries. Particular concerns were raised about the lack of technical standards for expressing rightsholder opt-out preferences. Opposition boiled over during the passage of the Data (Use and Access) Bill in mid-2025, where the House of Lords sought to insert provisions on transparency for copyright owners regarding the use of copyrighted works for AI training. Those transparency proposals were rejected, but as a concession to enable the Data Bill to pass, the government committed to publish a report and impact assessment within 9 months of the Data (Use and Access) Act ("DUAA"). 

What does the government update address?

The topics required to be covered in the report and impact assessment were outlined in sections 135-136 of the DUAA. This included an obligation to consider and make proposals in relation to:

• addressing the copyright and AI policy options discussed in the 2024 Copyright and AI consultation document;
• technical measures and standards to control the use of copyright works to develop AI systems;
• the effect of copyright on the access to data by developers of AI system;
• the disclosure of information by developers of AI systems about their use of copyright works in the development process; and
• the licensing of copyright works to AI developers for training and enforcement. 

The DUAA required the government to "make proposals" about a range of topics. A cynic might suggest that the government has instead only proposed to continue to monitor and consult. For example, on the issue of technical standards that enable rightsholders to express how their work can be accessed and used, the government concludes that:

"We propose to keep the need for regulation of technical tools and standards under review, and to continue to monitor international developments. We propose to work with experts and industry to support best practice and adoption of market-led tools and standards."

In addition to the topics mandated by the DUAA, the report also addressed section 9(3) of the Copyright Designs and Patents Act, which provides protection for literary, dramatic, musical or artistic works which are computer-generated; and the extent to which existing law adequately addresses deepfakes or digital replicas - the use of AI to replicate or mimic the appearance or voice of individuals. Again, the report does not reach a concluded policy provision on either of these issues. On section 9(3) the government intends to continue to monitor the use and impact of the provision, but indicates that it may repeal it. On digital replicas, the government identifies support for enhanced protections for a person's image and voice, but that there is no single view on what form that protection should take, and commits to launching a consultation on the issue of digital replicas in summer.

What next?

The issue of the relationship between copyright and AI is complex, but the UK creative industry will be disappointed that, despite considering the issue for over a year, the Government has been unable to meaningfully set out firm legislative reform proposals to address the pressing issues identified in the DUAA. 

The creative sector will welcome the conclusion reached by the Government in the Report that various stages of the development and deployment of AI models are likely to infringe copyright law if carried out in the UK without permission of copyright owners. 

While the Impact Assessment notes that "[s]ome aspects of UK copyright law and its application to AI training are disputed", the main source of uncertainty over the past year can arguably be traced back to the Government itself - specifically, its past approach of setting out a 'preferred option' reform proposal that would have weakened UK copyright law. The Report acknowledges that the legal position in the UK in relation to copyright and AI is one of several "uncertain factors" which has affected the extent to which the licensing market has developed; and consultation respondents highlighted that uncertainty created by the Government proposals for change "has a direct adverse impact on the current market for rightsholders". 

Despite this, the Government has not definitively ruled out the introduction of a new TDM exception with an opt-out reservation of rights mechanism, as the House of Lords Communications and Digital Select Committee urged it to do last week. Ruling out such a TDM exception would place UK copyright owners in a stronger position when negotiating licensing deals with AI developers, but for now rightsholders have been left to fend for themselves under existing principles of UK IP law – with government indecision looming large in the background.