The UK Government has published its long awaited Report on Copyright and Artificial Intelligence, following a consultation that attracted more than 11,500 responses. We covered the consultation in a previous blog – see here.
At the heart of the issue is whether AI companies should be able to train their models on copyright protected works without being transparent about the data used and without licensing or paying rightsholders.
As expected, the consultation highlighted a stark divide. Creative industry respondents overwhelmingly favoured stronger legal protections, particularly licensing requirements for AI training and greater transparency around AI outputs. Technology sector respondents, by contrast, generally supported more permissive rules to facilitate AI growth and development.
Despite the scale of engagement, the Government has stopped short of proposing legislative reform at this stage. Instead, it signals a period of consolidation, further evidence gathering and close monitoring of court decisions and overseas developments.
Striking a Balance
The Report frames the Government’s approach as a balancing exercise between two strategically important sectors:
“We want to ensure that our approach to copyright allows us to realise the extraordinary potential of AI to grow the economy… while protecting the UK’s position as a creative powerhouse.”
In practice, however, the Report largely preserves the status quo. Difficult policy choices are deferred while the Government waits for developments, such as the forthcoming appeal in the much-commentated-on Getty Images v Stability AI case, as well as legal and regulatory trends in jurisdictions like the United States. Technical standards are to be left to industry led initiatives rather than legislation.
One might say that, having climbed onto the fence, the Government has decided it is a comfortable place to sit, at least for now.
Text and Data Mining: A Retreat from Reform
The most significant outcome is the Government’s retreat from its earlier preferred option for a broad text and data mining (TDM) exception, which would primarily have benefitted AI developers.
The consultation considered four options:
- Retain the status quo;
- Strengthen copyright, requiring licensing in all cases;
- Introduce a broad TDM exception; or
- Introduce a TDM exception with opt out and transparency measures.
The consultation had suggested that the Government favoured option 4, broadly aligned with the EU’s approach. This would have allowed AI training on copyrighted works unless rightsholders opted out.
That position has now been abandoned following strong opposition, particularly from the creative industries. The backlash coalesced into coordinated public action, including a “silent album” released by artists such as Kate Bush, Annie Lennox and Damon Albarn, symbolising the risk that unlicensed AI training could effectively “silence” human creators, alongside the national “Make It Fair” campaign in which major UK newspapers ran identical front pages to highlight concerns over the erosion of copyright protections.
Acknowledging the criticism of option 4, the Government has seemingly (though not explicitly) pivoted towards a licensing-first model, coupled with improved transparency. The Report acknowledges a clear consensus that licensing should remain central, with the options of collective licensing through CMOs and direct licensing from rightsholders, being considered.
The Government will also continue to monitor international approaches. Japan, often cited as a permissive example, in fact applies a relatively narrow TDM exception, limited to non expressive uses that do not unreasonably prejudice rightsholders’ interests.
Transparency: Broad Support, with Proportionality Caveats
Transparency emerged as an underlying principle commanding broad support. Over 90% of respondents supported obligations on AI developers to disclose training data sources, alongside clearer technical standards to support compliance.
The Government agreed with this position in principle but stressed the need for proportionality, particularly to avoid imposing unreasonable burdens on SMEs and individual developers.
There was also strong support for transparency around AI outputs. Mandatory labelling of wholly AI generated content was widely favoured. Views were more divided where AI is used merely as an assistive tool in human creative processes.
For businesses developing or deploying AI systems, this points towards an eventual compliance obligation rather than immediate legal risk but, directionally, signals where future regulation is likely to land.
Computer Generated Works: Likely on Borrowed Time
The Report also addresses the UK’s distinctive protection for Computer Generated Works (CGWs), a regime with few international equivalents.
Introduced in the 1980s, CGW protection sits uneasily alongside modern originality standards for copyright that emphasise human creative input. In the context of generative AI, its boundaries remain legally untested.
Those who responded to this part of the consultation overwhelmingly supported removal of CGW protection, citing legal uncertainty and little evidence that it incentivises creativity or innovation. The Government appears persuaded, noting that copyright’s primary purpose is to reward human creativity and that CGWs are rarely relied upon in practice.
Other Issues: Enforcement, Standards and Digital Replicas
The Report also touches on:
- Enforcement: continued collaboration with industry, law enforcement and the courts to ensure effective remedies for AI related infringement;
- Technical standards: best developed through industry led initiatives with Government support, including standards applicable to web crawlers and AI developers; and
- Digital replicas: growing concern over AI generated voice and image replicas, particularly given the UK’s lack of standalone image or voice rights. Proposals such as a new “personality right” are likely to attract further attention.
Where Does This Leave Us?
For now, the UK remains in policy stasis. The Government has stepped back from controversial reform but has offered no definitive answers on what the solution is to this important issue. Both rightsholders and AI developers will need to watch litigation, particularly the Getty Images v Stability AI case, and international developments closely.
While the Report reassures creative industries that a broad TDM exception is now off the table, it prolongs uncertainty for businesses seeking legal clarity. In contrast, the House of Lords has taken a much firmer stance in its own report on the issue - rejecting a commercial TDM exception with opt out and calling instead for a licensing first approach, stronger transparency obligations and robust AI labelling. It warns that continued Government indecision risks stalling the development of a functioning licensing market altogether, arguably making the current policy limbo increasingly hard to justify.
For now, the Government has not taken a definitive position; however, this continued inaction is contributing to reduced certainty and may be constraining growth, with the risk of more difficult consequences when a decision is eventually made.
Please get in touch with any of the authors of this blog if you would like further information.
All of our thought leadership content relating to AI can now be found in our AI Hub – see here.
