The Copyright and AI Consultation
The UK Government’s 10-week Copyright and AI Consultation finished on 25 February 2025. The consultation sought guidance on plans to address the legal uncertainty surrounding Artificial Intelligence (AI) and intellectual property law.
The present situation sees rights holders facing difficulties with controlling the use of their copyrighted works in AI training, resulting in what some would see as inadequate and unfair compensation. On the other hand, the legal ambiguity for AI companies may be compromising investment, innovation and adoption of AI in the UK, thereby threatening the UK’s competitive standing in global development.
To manage the rising tensions, the UK Government launched a consultation in December 2024 inviting relevant parties to comment on proposed amendments to the current copyright legislation.
The New TDM Exception
A key point of consideration is the law relating to text-and-data mining (TDM), a process used by AI models to copy, extract and analyse large quantities of potentially copyrighted online material. The current UK position, set out in section 29A of the Copyright, Designs and Patents Act (CPDA) 1988, permits copying of a work for TDM as an exception only for non-commercial purposes.
The four options proposed in the Government Consultation are as follows:
- Option 0: Do nothing.
- Option 1: Strengthen Copyright.
- Option 2: Introduce a broad, mandatory TDM exception.
- Option 3: Introduce a TDM exception giving rights holders the chance to opt-out, supported by transparency measures.
The Government initially supports option 3, which would enable the mining of all works that are freely available or available through subscription on the internet, for both commercial and non-commercial purposes. This option would offer rights holders the chance to “reserve their rights”, but only by opting out of the TDM exception.
The Opt-out System
The consultation considered what such an opt-out system might look like for rights holders wanting to restrict use of their copyrighted material in AI training datasets. As such, the Government requested views on the current technologies used to regulate web crawlers including the robots.txt standard, metadata-level instructions, and do-not-train registries.
The consultation highlights several limitations of the above technologies and acknowledges the need for further development and standardisation both in the UK and internationally. Such limitations have not yet been resolved upon implementation of the EU TDM exception, leading to concerns about the Government’s ability to create a sufficiently robust opt-out system.
Transparency Measures
The consultation recognises the need to accompany any such legislative change with appropriate transparency measures, requiring sufficient disclosure from AI companies about the material used in the training of AI models, web crawlers used to mine training data, and measures taken to guarantee compliance with the above opt-out systems.
However, the practical reality of imposing such measures may prove challenging considering the sheer quantity of data used in AI training. In this regard, the consultation sought advice on how the UK might achieve sufficient transparency standards as well as views on the methods deployed in other jurisdictions.
Encouraging AI Innovation in the UK
The proposed broadening of the TDM exception would certainly clarify the UK’s approach, whilst aligning it more closely with that of the EU’s Digital Single Market (DSM) Copyright Directive (Directive (EU) 2019/790). This harmonisation could streamline compliance for companies operating internationally and might help to mitigate the administrative burden for rights holders across both jurisdictions. The looser TDM restrictions aim to incentivise AI companies to operate in the UK while attracting further investment and development opportunities from abroad, in line with the Government’s pledge to “make Britain the world leader” in AI.
Pushback from the Creative Industry
The Government’s proposal has been met with backlash from rights holders, primarily amongst those in creative industries, with concerns that allowing the “scraping” of copyrighted works by default risks harming the UK’s prosperous creative sector.
Alongside the widespread concern about the practical application of the Government’s proposal, one viewpoint considers the power imbalance that it might create between the parties involved. Current UK law places the obligation for acquiring lawful access to copyrighted material on the parties seeking that access. The proposed changes would shift responsibility onto the rights holder. For individual creators with varying resource levels and technical expertise, this could pose a significant challenge. Indeed, the administrative and logistical burden involved in reserving rights for large, complex portfolios of copyrighted works across numerous platforms gives rise to the possibility of material falling through the net.
Training Models using Pirated Databases
Despite issues surrounding the potential TDM exception, the need for legislative change is undisputed by all stakeholders alike. Recent allegations that Meta have trained their AI model, LLaMA, on the pirated database Library Genesis (LibGen), suggest that perhaps copyrighted works are at risk regardless of the existing laws in place. Unfortunately, Meta is not alone in these accusations, with numerous class action lawsuits underway against several other tech giants including, amongst others, OpenAI, Microsoft and Anthropic.
Conclusion and Future Perspectives
Given that many of the other leading global players in AI technology have already established their regulatory frameworks, the consultation represents an important and positive step forward for the UK. The proposed expansion of the copyright exception to include commercial TDM (albeit accompanied by an opt-out mechanism and increased transparency measures) unquestionably reflects the Government’s stated position on promoting AI growth, a position set out in detail in the recent AI Opportunities Action Plan.
As discussed above, the success of the proposed approach in practice will depend on the strategies the UK can develop to address the shortcomings highlighted by the Government and interested parties alike. It will certainly be interesting to see the Government’s response to those interacting with the consultation once the next report is published.
The above provides general commentary only and does not constitute specific advice. If you have any questions about any of the topics raised in this report, please contact us, or call us on 01206 571187 to speak to one of our attorneys today.
Image generated by Craiyon.
