What next for UK copyright law in the age of AI?

When I sat down to write this article, I began by asking my favourite Generative AI model (which I shall refer to as “GenAI”) whether it was trained on copyright works. It gave a cagey response:

“[GenAI], like other AI models, is trained on a mixture of licensed data, data created by human trainers, and publicly available data.” 

I followed up this initial query by asking whether scraping publicly available copyright works from the internet to train AI was lawful. GenAI responded with this:

“The legality of scraping publicly available copyrighted works from the internet to train generative AI models is a complex and evolving issue.”

True, but I realised that I had asked the wrong question. I tried again, typing: “do you need consent from the rights holder to train AI using copyright works under UK law?” At that point, to my surprise, GenAI conceded that:

“… while there are some provisions that might allow the use of copyrighted works for AI training under certain conditions, obtaining consent from the rights holder is generally required. Consulting with a legal expert is advisable to ensure compliance with all relevant regulations.”

GenAI is correct. In the UK, copyright gives the creators of an original work an exclusive right to reproduce that work, protecting against others copying that work without permission. Section 17(2) of the Copyright, Designs and Patents Act 1988 explains that:

“Copying in relation to a literary, dramatic, musical or artistic work means reproducing the work in any material form. This includes storing the work in any medium by electronic means.”

It’s generally assumed that training AI models involves reproducing works of this kind, infringing the creator’s copyright. The next question is whether the law strikes the right balance between the interests of AI developers and the interests of rights holders. In other words, does it need to be updated? At the risk of oversimplifying, there are two schools of thought on this in the UK:

1. The AI lobby: The law must change, otherwise we risk undermining economic growth by strangling the AI industry at birth.
2. The copyright lobby: The law must remain as it is, otherwise we are legalising theft and undermining economic growth by destroying the UK creative industry. 

In the UK, the Government is proposing a change to the law, introducing a data mining exception to copyright and requiring developers to publish details of the works used to train AI models. In theory:

• The data mining exception would help AI developers by making it lawful to use publicly available copyright works to train AI unless rights holders have explicitly opted out.
• The transparency requirements would help rights holders, as AI developers would be forced to open up the “black box” of training data on which their models are based. This would make it easier for right holders to enforce their copyright and generate licensing revenue.

Baroness Kidron (director of Bridget Jones, the Edge of Reason) described this proposal as “a forced marriage on slave terms” in a House of Lords debate. UK creatives – including the likes of Paul McCartney – have united behind this view, launching a series of publicity stunts, including: 

• The “MAKE IT FAIR” newspaper campaign.
• The release of a silent album backed by the likes of Kate Bush, Jamiroquai, Annie Lennox and Billy Ocean with the following track list: The, British, Government, Must, Not, Legalise, Music, Theft, To, Benefit, AI, Companies.

Meanwhile in the US, Sam Altman and OpenAI appear to be heading up the AI lobby, responding to a consultation on the development of a US “AI Action plan” with this:

“Applying the fair use doctrine to AI is not only a matter of American competitiveness —it’s a matter of national security. The rapid advances seen with the PRC’s DeepSeek, among other recent developments, show that America’s lead on frontier AI is far from guaranteed. Given concerted state support for critical industries and infrastructure projects, there’s little doubt that the PRC’s AI developers will enjoy unfettered access to data—including copyrighted data—that will improve their models. If the PRC’s developers have unfettered access to data and American companies are left without fair use access, the race for AI is effectively over.”

Comment

The UK Government’s AI and copyright consultation closed on 25 February and it will be interesting to see where it leads. To borrow Baroness Kidron’s colourful metaphor, the current signs indicate that the Government plans to hold a shotgun wedding between copyright and AI by pushing through its proposed changes to the existing law. Speaking at a conference in Silicon Valley last month, UK Technology Secretary Peter Kyle urged the copyright lobby not to “resist change”, noting that:

“We have the third-largest AI market in the world and we have the second-largest creative arts sector in the world. I will not pit one against the other and I will not be made to choose between one or the other.”

While there is a logic to the Minister’s position, it’s difficult to see how copyright law can be amended without making some politically difficult choices. There will be winners and losers – and time will tell whether it is the AI or creative industry that draws the short straw. 

Our content explained