On 18 March 2026, the Technology Secretary, Liz Kendall, made a written statement to Parliament confirming a significant reversal of government policy on copyright and artificial intelligence. In December 2024, the Government published a consultation on copyright and AI ,which we previously covered here, with its preferred approach being to enable AI developers to train on copyrighted works while giving rights holders the ability to opt out of the regime. It was termed as a "data mining exception with opt-out and transparency measures." The proposal would have seen a fundamental change in the way in which copyright could be exercised significantly reducing the scope of challenges to unauthorised use of copyright by AI developers. The proposal was overwhelmingly opposed by the vast majority of stakeholders in the creative industries.
The reversal marks a significant climb-down. Although the plans were not properly and fully outlined, in practice, they would have amounted to a carte blanche for the use of copyrighted material without permission with the possibility for the copyright holder to register dissent. Despite the UK Government's position that the proposed plan would "remove legal ambiguity", the practicalities of this arrangement were largely unclear, especially in terms of:
- what steps the copyright holder would have had to take to make sure they have opted out effectively;
- what was to happen where an AI tool was already processing copyrighted material, but a copyright holder registered an opt-out after the event; and
- the enforcement of copyright with AI developers (dispute resolution mechanisms) and access to information in relation to the content and data used to develop AI systems (in the absence of ironclad transparency measures)
In light of these difficulties and others, it is not a surprise that the UK Government has sought to de-label this as its preferred approach. The plan had also attracted major backlash from prominent artists, including Sir Elton John, who compared the use of copyrighted works without permission to "committing theft, thievery on a high scale."
The Law to Remain Unchanged
The policy reversal preserves – for now – the full force of the existing copyright framework under the Copyright, Designs and Patents Act 1988, under which rights holders retain the exclusive right to authorise or prohibit reproduction of their works, including for AI training purposes. Had the original opt-out regime been enacted, it would have constituted a profound retrenchment of IP protections: rights holders' primary cause of action for infringement of the reproduction right would effectively have been extinguished by default, leaving them dependent on successfully navigating an opt-out mechanism to preserve their rights. The abandonment of this proposal means rights holders' entitlement to bring infringement claims under the existing law remains unchanged. That being said, proof of such claims remains a significant hurdle given the complexity of data mining and processing.
Next Steps: Working Groups, Taskforce, and Ongoing Review
With the plan being abandoned, the Government's position is now without a clear alternative: it says it "no longer has a preferred option" for what to do next. The announcement confirmed that the Government would not reform copyright laws "until we are confident that they will meet our objectives for the economy and UK citizens."
In place of the abandoned opt-out regime, the Technology Secretary set out a four-strand work programme:
- The Government will launch a consultation in the summer on "digital replicas" – instances where someone's likeness is used without permission – while protecting legitimate innovation.
- A taskforce will be established to put forward proposals on best practice for labelling AI-generated content, with an interim report due in the autumn of 2026.
- A review of mechanisms available for creators to control their works online will be published, covering technical solutions and standards for input transparency.
- A working group will be launched specifically focused on independent and smaller creative organisations, to explore whether government has a role in supporting their ability to license content.
The Government also confirmed it is establishing a Creative Content Exchange – a trusted marketplace for digitised cultural and creative assets – supported by funding from UK Research and Innovation and sponsored by the Department for Culture, Media and Sport, with a pilot phase already under way.
Significant questions remain unanswered, including how creators would be compensated for past or future AI training use, and how transparency over training data would be enforced in practice. The Government noted that it is not the only administration grappling with the issue, pointing to the United States, the European Union, Australia, and India as fellow jurisdictions seeking a pathway through.
But while the UK is trying to chart a way through the issue, AI tools continue to operate, crawling through vast amounts of copyrighted material, in many instances without clear consent of the rights holders. It is unclear why the December 2024 consultation could not have been used to address the information lacuna the government now seeks to address with the establishment of various further consultations and working groups. While it is certainly true that other jurisdictions grapple with the same issue, the European Union was the first to enact a comprehensive AI law, namely the EU AI Act which was passed in August 2024. Further delays will undoubtedly impact both rights holders and AI developers, creating significant uncertainty.
