On 25th February 2025 many of the UK’s major newspapers featured the same front covers. Why? To protest against the government’s consultation on a possible exception to copyright infringement to allow tech companies to freely use copyright material to train AI models.
Ministers say the proposals aim to ‘protect the interests of our creative industries and AI sectors’, but they have attracted widespread criticism from across the creative industry. An open statement signed by over 47,000 creatives, from Kevin Bacon to Kate Bush, reads:
“The unlicensed use of creative works for training generative AI is a major, unjust threat to the livelihoods of the people behind those works, and must not be permitted.”
The issue is an outdated copyright regime which has not kept pace with the development of new technologies. To develop AI programs, access is required to underlying data and materials to “train” the programs to spot patterns and respond to user input.
In doing so, the developers reproduce that underlying data and materials as part of the technological process and, sometimes, in the resulting output. Under current law, it is likely that such reproduction without appropriate licences would infringe copyright.
Ministers point to the creative industries and AI sector as ‘UK strengths’ and vital to a ‘national mission’ to grow the economy. As such, it’s clearly imperative to get this right for both parties.
The groundswell of concern on the side of creative artists is the protection of their works – as a £126 billion a year industry, you can understand why. They want to see improved transparency around training data, enabling them to identify the use of their works.
On the other, AI developers are finding it difficult to navigate copyright laws, leading to tensions between the sector and current legislation. Ministers fear this will threaten future investment in and the adoption of AI technology.
The sector is also hugely valuable in the UK, with the latest government data suggesting that there are more than 3,000 AI companies, generating more than £10 billion in AI-related revenue, while employing more than 50,000 in AI roles. The stakes are high.
So, what’s currently on the table? Why does it fall short of expectations? And what is the solution?
Current proposals
The Department of Science, Innovation and Technology and the Department for Culture, Media and Sports have stated they have three objectives:
- Supporting right holders’ controlof their content and ability to be remunerated for its use.
- Supporting the development of world-leadingAI models in the UK by ensuring wide and lawful access to high-quality data.
- Promoting greater trust and transparencybetween the sectors.
The inherent conflict between these objectives and the industries involved has made it hard to agree a workable solution before now.
The government’s proposed solution is a copyright exception for ‘text and data mining’ (TDM) which is broadly in line with the EU position under the Copyright Directive 2019. It is also in line with proposals of the previous Government which were dropped in 2023, after an uproar from the creative industries.
Copyright owners would be offered the opportunity to reserve their rights and opt out of the exception. This would enable those rights holders to still demand compensation for use of their works for training large language models.
However, the concern for artists is that in doing so those works will not receive the same visibility as other works used by the AI programs and that the AI will compete with the original works.
Where are the gaps?
There’s no doubt that this a difficult and complex situation, where one side feels their creative work is fair game, and the other is feeling hamstrung by the current system.
Critics say that there has been no objective case made for a new copyright exception, nor has a water-tight rights-reservation process been outlined anywhere around the world.
There is also a question mark about whether the exception would be retrospective and cover past infringements by AI developers for programs that are already on the market.
The solution?
As the government itself admits: “Delivering this will be challenging and will require practical technical solutions as well as a clear legal framework. It will require us to work closely with our international partners to progress towards an interoperable AI and copyright framework.”
The exception would automatically apply to any works which the user can lawfully access (including from behind a paywall where the developer has paid for access).
One concern for authors and artists is how the opt-out can be enforced and so the government is looking at technical solutions to achieve enforcement. Further concerns relate to transparency and whether AI developers would be open about the sources they have used to train their programs.
It seems inevitable that something will change, not least because there is concern the UK will be left behind in the AI revolution otherwise. How far the government is prepared to go remains to be seen and we can expect further significant pressure from copyright owners in the meantime.
