The UK’s new AI-copyright consultation: a path to clarity or controversy?

The UK government has just artificially started an inquiry into proposed copyright law reforms. It seeks to get to grips with the complex and often contentious interplay between AI and IP rights. This consultation focuses on matters such as TDM, the transparency of AI training, and the legal status of AI-created works. These measures, which on the one hand aim at providing a more clear legal framework, also spark a debate about the balance between innovation and rights of content creators.

This guide will make you go over the main aspects of the topic and help you to understand the most important information.

Expanding the TDM Exception

By training AI models, vast amounts of data are needed – most of it copyrighted. The main controversy lies in whether or not the extraction or utilization of such data works without express consent from the rights holder. The best way to balance the competing interests of AI developers and content owners remains at issue. TDM for Commercial Purposes: An issue of licensing. The current CDPA only allows TDM for non-commercial research. The government is currently contemplating an extension in the law that would allow TDM for commercial goals, by this means facilitating users to extract data from any material to which they have lawful access, whether openly available on the Internet or behind a paywall. In that case, “reserve their rights” can be done by the rights-holders opting out, thereby requiring the AI developers to go through a licensing process. This proposal seems to be more in line with the DSM Copyright Directive within the EU.

Opinions are invited on whether this is the best way to go and whether alternative solutions may have to be retained, for example, maintaining the status quo, mandating express licensing in all circumstances, or a broad TDM exception without an opt-out mechanism being constructed. Further views are sought in relation to the technical means for the right holders’ exercise of opting out in order to avoid ambiguity with regard to the implementation of the EU; the following examples would be included:

• Robots.txt: Very common protocol that informs AI crawlers about the areas of a website they can visit;
• Metadata: Embedding rights to use a work in its metadata.
• Notification of AI developers and dataset owners directly, potentially through a third-party register.

Each has its drawbacks, whether in the form of a lack of finely grained control over individual works, possibly low uptake of metadata integration, or doing direct notifications at scale. The government is looking at a way forward which is practical and workable without adding inordinate regulatory burden.

Ensuring Greater Transparency

The UK government believes that transparency is essential concerning the stated changes. They said an exception for TDM would only come about with robust disclosure demands. Most concern lies in the fact that developers of AI surprisingly do not unveil their sources while training. Without further disclosure, rights holders will still not be able to ascertain whether their works were used, thereby making enforcement difficult and increasing uncertainty about liability.

However, enforcing clearance rules presents practical challenges due to the sheer volume of data involved, alongside legitimate concerns about trade secrets. Feedback is requested on how to strike a fair balance, with the EU’s model serving as a possible framework. The EU requires AI developers to publish a “sufficiently detailed summary” of their training data, without mandating exhaustive disclosure. The UK appears inclined to follow this approach but is open to refining it further based on stakeholder input.

The Legal Status of AI-Generated Works

A significant point of debate is the copyright status of content created entirely by AI. Section 9(3) of the CDPA currently assigns ownership of computer-generated works without a human author to the person who made “the arrangements necessary” for their creation. The government has identified two primary concerns with this provision:

1. It contradicts the principle that copyright requires human originality, expression, and creative intent.
2. It may be unnecessary, as AI-assisted works where a human plays a role in creation are already protected.

As a result, the government is considering removing this provision entirely but is seeking input on whether it should be retained, modified, or abolished. Additional feedback is being requested on whether AI-generated works should be subject to mandatory labeling and whether existing licensing agreements provide adequate protection for original creators whose works are used in AI training.

Will These Changes Bring Clarity or Spark Further Debate?

The UK’s proposed reforms aim to create a clearer legal landscape. However, they also introduce new uncertainties. AI developers may benefit from expanded TDM exceptions, while rights holders may argue that an opt-out system does not offer enough protection. Transparency measures could foster trust, but implementing them fairly and efficiently remains a challenge. Meanwhile, the conversation surrounding AI-generated works raises fundamental questions about authorship and creativity in the digital age.

As discussions progress, the outcome of this review will shape the future of AI and copyright regulation in the UK. Whether it resolves existing ambiguities or leads to new controversies remains to be seen.

Conclusion

The proposed reforms seek to bring much-needed clarity to a rapidly evolving landscape, yet they also introduce fresh complexities. Expanding data-mining exceptions, enhancing clearance, and reconsidering the legal status of certain works could reshape the relationship between technology and creative industries. However, the challenge lies in striking a balance that fosters innovation while safeguarding the rights of content creators.