Industry bodies welcomed the government’s announcement last month that it has ditched plans to allow AI firms to use copyrighted works without permission.
Here, Gregor Pryor, managing partner EME at Reed Smith, looks at the developments that led to this point and what the industry should expect to happen next with AI and copyright…
After many months of uncertainty, the UK government has abandoned plans to allow the operators of AI models to use copyrighted works without permission. Whilst this stops some way short of a complete victory in favour of creators, it is a clear step towards a more rights-holder friendly national approach. It is welcome news for many in the music industry. But there are still big issues on the horizon.
To recap, the government has ditched the broad exception for text and data mining and controversial plans which would have seen music rights-holders having to actively opt-out of their copyrights being used by AI companies. Encouragingly, there is also a proposal to collaborate with AI companies to ensure transparency and best practice.
The report strikes a realistic balance by acknowledging the fast-growing market for legitimate licensing for AI training data, perhaps paving the way for an adaptable and sophisticated licensing ecosystem which benefits both creatives and technology companies moving forward.
Assuming that AI is clearly here to stay – how did we get to this latest report? And what could be next for the music industry as it grapples with this new force?
The build up: copyright, content, and controversy
The intersection between copyright law and data protection has never been more topical. Yes, the Data (Use and Access) Act 2025 gave the country something of a roadmap for navigating these challenges, but questions and controversies still abound. Only at the start of March did the House of Lords Communications and Digital Committee publish a report declaring that ‘UK creative industries face a clear and present danger from generative AI’.
For instance, data relating to music content typically contains both copyright and personal data in an AI context. The music and tech industries have been waiting for much-needed guidance on how these two legal protections will be dealt with. Licensing pilots are already underway in attempts to forge a path, but they are still nascent and more government guidance has been needed.
There’s also the political context. In 2023 Rishi Sunak detailed the risks and responsibilities which AI brings. Labour’s manifesto pledged to “ensure the safe development and use of AI models” whilst in the King’s Speech they aimed to “harness [its] power”. Following the pushback Labour’s proposals have received since – driven in no small part by music industry figures – March’s announcements were also an attempt to win back favour and push ahead with policymaking.
Top recommendations
The announcements raised several proposals for the music industry to digest – some were more significant and surprising than others.
The biggest news for the music industry is the removal of the opt-out mechanism. This has been the source of much previous controversy – and concern – with the likes of Elton John and Dua Lipa leading the criticism. Although the wider proposals may have their own strengths and weaknesses, this climbdown by the government is a clear win for music. It is only logical that rights-holders are the ones by default who decide how their content is used, and the advent of AI shouldn’t change this.
It is only logical that rights-holders are the ones by default who decide how their content is used, and the advent of AI shouldn’t change this
Gregor Pryor
Second, the government has suggested repealing current provisions in UK copyright law which afford protection to computer-generated works. In many respects, the UK was an outlier when it came to protecting AI-generated works, compared to laws in many other countries. As has happened in the US, this type of change may give rise to questions concerning how critical human involvement is for determining if a work should be protected by copyright.
For artists and songwriters who include AI-generated elements in their work, how will this proposal be put into practice and, more importantly, will it affect their wider protections?
Third, several ideas were put forward for dealing with digital replicas and deepfakes, including a somewhat radical suggested personality right. It’s hard to find a record label or artist that isn’t against deepfakes. With replica and soundalike songs already rife across streaming platforms, the industry will be tracking which proposals are pursued and how much protection they ultimately give.
Overall, the government is clearly willing to trial a largely free-market approach, seeing how the UK industry develops and how other countries react. Whilst some in the industry will welcome this more hands-off strategy, others will be concerned it will allow further threats to creativity and copyright to foment.
Name, image, and likeness rights – the next AI frontier for music?
Amid broader concerns, one area where collaboration between creatives and tech is thriving is avatar concerts. Following pioneering predecessors, ABBA Voyage has particularly taken audiences by storm. A KISS avatar show is set to debut next year, and we will almost certainly see others follow.
Name, image, and likeness rights are at the core of these projects. Our firm has helped shape the law for artists in this area, following the success of the artist Rihanna in her case against Topshop for using her image without permission. Still, image and personality rights are a hugely complex and under-developed area of law in the UK. Add to this topic the prospect of AI, and complexities intensify. For example, given that AI has the potential to massively extend the longevity of these rights, more questions arise around how they can be protected.
The report highlights that this intersection of copyright and data protection law is ripe with creative and technological opportunities.
Lessons from streaming?
For all the potential for innovation and change that AI can bring to the industry, there is still plenty of room for disruption. Many of us are old enough to remember that this is not the first time music has had to contend with dramatic changes caused by new technology. The prevalence of piracy in the late ’90s and during the dot com boom, as adoption of the Internet took hold, was a devastating period. It wasn’t until the rise of legitimate, licensed streaming and downloading that we turned the corner of opportunity.
Although AI poses more complex legal and philosophical questions than streaming, there are parallels. Several of the early digital music platforms were infringing, but nonetheless cleared the way for legitimate alternatives. Spotify sold itself in the early years as a way to combat piratical services in Sweden. In the case of LLMs, though, commentators disagree about the extent to which copies made in the course of training, and the act of training itself, are unlawful. Until the law settles globally, the industry must reckon with a technology that some say is obviously cannibalistic.
There are similarities in trajectories, too. The early days of legitimate downloading were dogged by resistance, with rightsholders repeatedly wanting to suppress new technologies until commercial and structural models were locked in. Conversely, technology companies in almost all cases wanted to charge ahead at pace – ‘move fast, break things’. Just as streaming companies eventually forged a path to legitimate licensing models, so now are LLMs genuinely exploring what legitimately trained AI models can look like. Although streaming ultimately proved impossible to live without, it remains to be seen whether AI will become truly part of the furniture.
The history of streaming cannot answer every question about AI nor assuage every concern. However, while we are deep in the weeds of developing AI laws, regulations, and industry practice, it gives a reassuring reminder that in the long run the music industry is more than capable of not only rising to the challenge of technology innovation, but thriving.
‘ The preceding article may include information circulated by third parties ’
‘ Some details of this article were extracted from the following source www.musicweek.com ’
