In December 2024, the Government issued a consultation on copyright and AI, which has been one of the most engaged-with government consultations in recent years, with over 11,500 responses by the time it closed in February 2025. Whilst the consultation considers a number of issues relating to the interplay between generative AI and copyright, it was the proposal to allow the use of copyright works to train AI models unless rights holders specifically opted their works out, that proved particularly controversial, leading to a sustained campaign in opposition, led by many leading players in the creative industries. This opposition also played out in Parliament, with the protracted Parliamentary ping-pong consideration of copyright-related amendments that arose during the passage of the Data (Use and Access) Act (DUAA) leading to a compromise requiring the Secretary of State for Science, Innovation and Technology to publish two key documents by 18 March 2026:
- An economic impact assessment of the options put forward in the consultation on AI training; and
- A report on the use of copyright works in the development of AI systems.
In the meantime, as also required by the DUAA, the Government has published its statement of progress on these issues, ahead of publication of the final report.
Consultation responses
The Government confirms that the UKIPO has now conducted a review of the just over 11,500 responses received to the consultation (NB: over 3,000 responses were based on template letters). This has clearly been a mammoth exercise for the UKIPO, involving around 80 IP officials and analysts reviewing the responses and identifying relevant themes arising from them.
On the key question of whether to introduce a text and data mining exception, the responses indicate that of those who responded through the online survey service:
- 88% expressed support for Option 1 – requiring licences in all cases
- 7% expressed support for Option 0 – making no changes to copyright law
- 3% expressed support for Option 3, the Government's preferred option – introducing an exception for text and data mining with rights reservation
- 0.5% expressed support for Option 2 – introducing a blanket text and data mining exception i.e., with no rights reservation
The creative industries also expressed strong support in their responses in favour of statutory transparency measures, whereas respondents from the tech sector showed mixed responses on transparency requirements.
Clearly, there is an overwhelming rejection of the Government's preferred option in the above headline figures. However, the Government notes that the distribution of preferences partly reflects the large response from individual creators and the creative industries generally. For respondents from the tech sector, options 3 and 2 were (unsurprisingly) the favoured proposals.
The final report
The final report, to be published by 18 March 2026, must consider each of the four AI training options put forward in the consultation, and must also contain the Government's proposals on the following issues:
- Technical measures and standards that may be used to control the use of and access to copyright works to develop AI systems.
- The effect of copyright on access to and use of data by developers of AI systems.
- Disclosure of information by developers of AI systems about their use of and access to copyright works.
- Granting of copyright licences to developers of AI systems.
- Enforcement of rules relating to use of and access to copyright works to develop AI systems.
Importantly, the report must also consider the treatment of AI systems developed outside the UK (a key issue following the High Court's recent decision in Getty Images v Stability AI).
Stakeholder engagement and working groups
Separately, there has been a programme of stakeholder engagement and working group discussions focusing on issues including:
- Control and technical standards – this group is exploring effective technical tools and standards (an issue that is also being considered at the EU level)
- Information and transparency – e.g. training data summaries, again an issue that has already been considered at the EU level
- Licensing – including the strengths and weaknesses of the current licensing framework
- Wider support for the creative industries
Comment
Opponents of the proposal to introduce a text and data mining exception will be buoyed by the overwhelming opposition to the Government's preferred option on AI training. The Government is clearly in no doubt as to the strength of feeling, and underscores the importance of ensuring the UK's creative industries continue to thrive in the AI era. However, it also highlights that it needs to act on its commitment to promote investment and development of AI as a key plank of national renewal and economic growth. The content of the March 2026 reports will therefore be keenly anticipated.
To track both case law and legislative/policy developments relating to GenAI and copyright, sign up to receive alerts from our tracker.