Labour’s Showdown on AI and Creative Rights

Robyn Dennis 
19/June/2025


After years of slow progress, collapsed drafts, and last-minute controversies, the UK’s Data Use and Access Bill has finally cleared its last hurdle and will receive Royal Assent today (19 June). For Labour, this is a legislative win, and a manifesto commitment ticked off – modernising the UK's data regime and clearing a path for smarter digital infrastructure across both public and private sectors. But for the creative industries, and those whose work is quietly fuelling AI, the moment feels less like a finish line and more like a warning flare.

The new legislation has been a long time in the making. First introduced by the previous Conservative government, it arrived in Parliament not once, but twice, before being unceremoniously dropped during the pre-election wash-up. Labour’s rebooted version, launched in October 2024, preserved much of the original structure: recognised legitimate interests, streamlined Data Subject Access Request exemptions, updated rules on automated decision-making, and more flexible legal bases for further processing. For many sectors, the most impactful elements are the new digital identity and customer data-sharing frameworks – cornerstones for future smart data schemes and cross-sector interoperability.

It is, on paper, a serious piece of technocratic legislation, providing a legal foundation to deliver more integrated, consumer-centric digital services. But it was never really the data-sharing provisions that set political pulses racing. That honour belongs to the rows erupted late in the Bill’s journey – over the use of creative works to train machine learning models.

At the heart of the controversy was a proposed provision that would have allowed AI companies to use copyrighted material for model training without permission or compensation, so long as rights holders were given a way to opt out. Critics, including musicians, authors and visual artists, called it a digital land grab. The issue snowballed from an obscure legislative clause into a full-blown political headache, with Sir Paul McCartney and Ed Sheeran on one side, and a nervy frontbench on the other.

Baroness Kidron, a longtime advocate for digital rights, introduced a series of amendments aimed at increasing transparency and strengthening protections for creators. This was designed to ensure rights holders could actually know whether their content had been used in the first place – thus making the opt-out meaningful. It was, in effect, a data transparency requirement dressed in the language of copyright. The amendment passed the Lords with strong support but was ultimately rejected in the Commons after five successive defeats.

Attempting to calm the storm rather than sail through it, Science, Innovation and Technology Secretary of State Peter Kyle announced the government would address the thorny issues of copyright and AI transparency in a separate, more comprehensive AI bill. But rather than settling the debate, the move appears to have simply kicked the can down a longer, more uncertain road.

Originally, the Labour government had planned to introduce a standalone AI regulation bill shortly after taking office. That legislation was set to include stringent provisions, including a requirement that high-risk models be submitted to the UK’s new AI Safety Institute for testing. But those plans were shelved in February to allow time for “international alignment” – shorthand, insiders say, for staying in step with Washington and concerns over the UK’s attractiveness to AI companies and investors.

Now, ministers are indicating the AI and copyright issues will be folded into a single, far-reaching legislative package in the next parliamentary session. In practice, that means nothing will land before the King’s Speech in May 2026 – effectively pushing the resolution of the copyright and transparency debate into the political long grass.

This delay has been met with frustration by creative industries and rights holders who were promised clearer protections. They now face the prospect of an extended period of legal uncertainty, with fears that both the timeline and the strength of the protections could be compromised. Observers warn that the desire for international consistency, while important, risks sidelining bold domestic measures that could have set a global benchmark for AI transparency and creators’ rights.

So, while the Data Use and Access Bill delivers on its promise of modernising the UK’s data infrastructure, it leaves the bigger battle unresolved. Who owns the training data that fuels artificial intelligence? What rights do creators have over the works they produce in an age of synthetic content? And how transparent should AI developers be about the data their systems consume?

In the Data Bill, Labour believes it has demonstrated its ability to pass complex, industry-focused legislation with competence. But on AI and copyright, it has chosen to delay decisive action. Whether this proves to be a clever political calculation, or a missed opportunity to establish global standards, will depend on the balance and timing of the forthcoming legislation.

For now, the infrastructure is in place, but the clock is ticking. Clear, fair rules on how AI uses creative works aren’t just overdue, they’re essential. Striking the right balance between protecting creators and fostering innovation will be key to ensuring the UK leads responsibly in the global AI landscape. The priority for Labour must be to establish a regulatory framework that evolves in quick step with technological advancements, providing clarity and confidence for creators, developers, and users alike.