Jamie Watt and Simon Casinader of the MARQUES Copyright Team report on the results of the UK government consultation on copyright and AI.

The consultation

From 17 December 2024 to 25 February 2025 the UK government ran a consultation upon copyright and artificial intelligence. The consultation considered four government specified policy options on copyright and the training of artificial intelligence models, as well as various alternative approaches.

The four options proposed in the consultation were for the government to:

1. do nothing, and allow existing laws to remain in effect as currently drafted in relation to AI and copyright;
2. require licensing of rights under copyright in all cases;
3. create a broad text and data mining exception, allowing significant use of copyright works in the context of AI systems without rights holder permissions; and/or
4. create a text and data mining exception but coupled with an opt-out and processing transparency requirements.

In addition, the consultation also considered:

• the effect of copyright on access to, and use of, data by AI developers;
• transparency regarding access to, and use of, copyright works by AI developers, and about the outputs of AI systems;
• technical measures and standards that may be used to control access to, and use of, copyright works to develop AI systems;
• licensing of copyright works for AI development;
• enforcement of requirements and restrictions relating to the access to, and use of, copyright works to develop AI systems, and relating to their outputs; and
• computer-generated works and digital replicas

It was noted that the consultation could result in revisions to copyright law, and measures promoting transparency, promoting use and concerning licensing of copyright works.

In March 2026 the results of the consultation were published. They can be found here.

Factors considered in compiling the results

In compiling the results, the government recognised the effects of technology in allowing for new ways in which copyright works may be copied and communicated to the public. They also recognised the need to balance rights provided to creators with the public interest in allowing for exchange of information.

Specifically, they identified that AI can allow for greater information analysis and availability, but that AI systems would not work effectively without the presence of the results of human creativity. They further identified that AI systems can in some instances compete with the very creators that they learn from.

The government also identified a need to protect the UK’s creative economies, and to ensure fair reward for uses of resources creators may create. They noted uncertainties around how copyright and AI interacted, particularly in the case where autonomous agents may be involved, the growth of licensing regimes, ongoing development of technical standards and the presence of ongoing litigation in various jurisdictions. Particularly noted were gaps in available evidence.

What the report on the consultation said

A copyright exception for AI training

Responses (particularly those from the creative industries) generally indicated a rejection of a broad text and data mining exception, including where an opt-out would be provided. Certain respondents in the AI and research sector felt that such an approach would not achieve the aim of making the UK internationally competitive.

Noting the strong views expressed in the consultation and gaps in available evidence, the government noted that a broad copyright exception with an opt-out was no longer their preferred way forward. Instead, they proposed to “gather further evidence on how copyright laws are impacting the development and deployment of AI across the economy…consider and engage stakeholders on other potential policy approaches…[and to] continue to monitor developments in technology, litigation, international approaches, and the licensing market”.

In choosing this approach the government noted the importance of enabling the transformational benefits of AI whilst balancing this with the need to protect human creativity and the UK’s creative industries.

Transparency

Noting that some countries require AI developers to disclose training data sources, and that the majority of respondents favoured a similar approach, the government noted their agreement, however they proposed only to “continue monitoring the effects of transparency rules in other countries and consider [their] approach in the UK”.

In this context reference was made to working with a range of industry and other experts to develop best practice on transparency, but in a way which was proportionate and which would not involve unreasonable burdens, but which would promote clarity and enforcement for stakeholders of all sizes, without disincentivising AI development or deployment in the UK (including by non-UK developers).

Labelling of AI and human-created content

The government identified that the labelling of content to achieve clarity in the use of AI in its creation can affect people’s choices and protect against disinformation and harmful deepfakes. They also identified that no labelling obligations currently exist in the UK, but that there was a “relative consensus” which existed amongst respondents in favour of the “principle of labelling”.

As a result, the government proposed to “work with industry to explore best practice on labelling AI-generated content… continue monitoring international developments and…work with international partners to support the development of common solutions”.

In this, they noted the role of best practices in helping establish consumer confidence and public trust in AI outputs. In furthering goals in this respect they identified the importance of complementing not duplicating global processes and common international solutions.

Technical tools and standards

The government noted that technical tools and standards enable rights holders to express how their work can be accessed and used, helping them control and license their works. They also noted that the market for these tools is developing rapidly, and that on a general level tool development was supported (although respondents differed as regards the nature of any intervention required).

As such the government proposed “to keep the need for regulation of technical tools and standards under review… to continue to monitor international developments…[and]… to work with experts and industry to support best practice and adoption of market-led tools and standards”.

Licensing

Acknowledging that creators’ works are widely used to train AI models, and the importance of creator remuneration (particularly in creator incentivisation), the government identified that AI could provide a new revenue source for creators. Respondents generally felt that the government should not interfere in the licensing market, but rather (particularly in responses from the creative sector) that transparency was paramount – as transparency requirements would enable rights holders to better license and enforce rights.

As such the government proposed to “not to intervene in the licensing market at this stage [but] to monitor the market as it develops and…keep market-led approaches to licensing under review”. They also proposed “with regard to AI systems developed outside the UK…to continue monitoring global developments and judicial outcomes” and “to identify and assess further levers to support access to valuable datasets”, particularly through the Creative Content Exchange (CCE).

The CCE is a government-established marketplace to sell, buy, license and enable permitted access to digitised UK cultural institution creative assets and is intended to test a range of commercial models for licensing, with the aim of launching an operational pilot platform by Summer 2026.

Also noted was the power of AI to transform education by helping teachers focus on what they do best. In this respect the report stated that the UK’s Department for Education will continue to take forward its work to ensure the copyright framework allows the benefit of AI to be realised in education settings, and is now assessing a range of options, including to ensure pupil IP is used fairly.

Enforcement

Acknowledging that effective rights enforcement mechanisms were key to ensuring rights remained meaningful, and that AI posed particular challenges, the government stated that enforcement should be effective, accessible and proportionate, and that it was important to ensure that the UK retained a competitive enforcement framework.

With this in mind they proposed to “continue working with partners, including law enforcement and the judiciary, to help ensure the UK enforcement framework remains fit for purpose…[and] to identify and address enforcement barriers and consider where action to mitigate these barriers may be required”. They also proposed “to consider the case for and approach to regulatory oversight of transparency or other measures, if legislation across these is introduced”. However they did not propose the establishment of any new “AI regulator” at this time.

Computer-generated works

The results acknowledged that in the UK the possibility for protection for computer-generated works created without a human author already existed, under present legislation. Most respondents felt that works solely generated by AI should not attract copyright, and that this part of UK law should be amended accordingly.

The government identified a need to incentivise and protect human creativity, and with the responses in mind proposed a balance in removing this specific type of protection whilst allowing copyright to continue to protect works created with AI assistance.

Digital replicas

The government identified that AI made it easier to create digital replicas of someone’s voice or face, and that harm could be suffered where this took place without permission. They noted the presence of some legal protections, but that they did not cover all situations where a replica may be made without consent, and the presence of some support for enhanced protections (but that there was no single view in this respect).

They agreed that AI presented increasing risks for artists and the public and proposed “to explore a range of options for addressing these risks, while protecting the potential of this technology to support legitimate innovation [including] consideration of whether it would be beneficial to introduce a new digital replica or personality right”.

Specific actions

The report noted various specific actions to take place following the release of the report. These included:

• Monitoring important developments in the wider global market, particularly in the US where most AI models are trained, and as regards transparency rules in the EU and parts of the US, and new technical standards under development.
• Using commissioned research to understand the impact of copyright reform on the creative industries, and building the evidence base on the supply and demand of creative content for use in training, to help build a framework for quantifying the impacts of copyright reform on the creative industries and AI sectors.
• Evidence gathering on the effectiveness on protection tools and standards and how these support the creative industries, competitiveness and access to data.
• Collecting evidence on developers across the tech stack and the extent to which copyright affects business activity, focussing on areas of UK strength such as fine-tuning, application-layer development, and supporting this via stakeholder engagement, including with sector and industry bodies, to understand how developers source training data and deploy text and data mining solutions in practice, and how they may respond to different policy options.
• Compiling analyses and evidence on AI’s impact on the economy and labour market, tracking the pace and pattern of AI adoption across sectors, as well as surveying the extent to which copyright affects AI adoption.

Comment

Whilst the proposals are in many respects a statement of continuing to wait and see how things develop, that approach may be considered sensible with respect to a fast-moving market where things will change and develop rapidly over the next few years. Legislating at this time may not be sensible, given the market may very well differ by the time any legislation is drafted and enacted. The recognition of fitting in with global approaches is perhaps to be welcomed as the UK, its rights holders and its developers do not exist in a vacuum.

Whether this will ultimately be beneficial for rights holders remains to be seen; indeed whether a unified global approach is possible may also be perhaps somewhat questionable. Given the prevalence of development located in other countries – and given the multitude of US litigation and their notable differences in copyright law when compared with the UK, and the relevance of the territorial location processing and communication in recent UK litigation – one may consider that the UK’s approach could simply devolve to mirror that of their most important economic partners. However, whether this will be the case is simply guesswork at this point in time.

What is known is that the UK’s approach remains fluid, to an extent. Where some certainty appears to be creeping in can be seen in relation to specific key points, as follows. It is highly unlikely that the UK will enact a broad text and data mining exception to copyright, whether in a standalone form or coupled with an opt-out mechanism. In the meantime, existing restrictions and approaches, and enforcement mechanisms, will continue to apply. We may, however, see some change in relation to the UK’s approach to allowing protection of computer-generated works to exclude works generated wholly without human intervention. We may also see some sort of greater protection for image rights, in the context of AI in the UK.

For rights holders, this should be generally welcomed. The key message, though, for rights holders to take from this is the importance of engaging with their industry and trade bodies, to address specific issues and further the development and implementation of industry standard approaches/non-legislative rights protection mechanisms and controls.

There was significant mention in the report of the importance and role of such bodies, and their relevance in shaping the rights landscape in the UK in relation to the development and use of AI should not be underestimated. They are also likely to play a key role in subsequent evidence gathering.

Industry bodies could also have significant impact on rights holders’ engagement with developers and the creation of balanced approaches in rights holder/developer relationships, acknowledging the fact that one needs both developers and creators properly incentivised and remunerated to make AI work.

Jamie Watt is a Partner and Head of the IP/IT team at Harper Macleod in Edinburgh and Simon Casinader is a Partner in the IP practice at K&L Gates in London