On 18 March 2026, the UK government published its long-awaited Copyright and Artificial Intelligence report (the "Report"), together with an associated economic impact assessment, in compliance with the deadline mandated by the Data (Use and Access) Act 2025. 

The Report seeks to address the complex relationship between copyright law and the development and deployment of AI systems, particularly with regard to the use of training data for generative AI models. It comes as the latest instalment in a debate that has occupied government policy over several years and, more recently, finds itself as the subject of copyright proceedings before courts in many jurisdictions, including in the UK in the Getty Images v Stability AI proceedings (though ultimately the relevant issues in that case fell away before judgment). 

The issue has been ripe for consideration even from the very early days following the initial introduction of the "text and data mining" ("TDM") exception into UK copyright legislation in 2014, when AI was more readily referred to as ‘machine learning’, but was advancing apace, nonetheless. When EU member states (though not the UK) implemented the DSM Directive in 2019 while the UK prepared for Brexit. In 2022, when the UKIPO launched a first consultation, which saw the creative industries and AI developers each share their views on the subject, and then again in 2024 when the Government launched a further consultation which outlined four policy options on the future of copyright and the training of AI models, amongst other things, which elicited some 11,520 responses.  

Perhaps as an indication of the significance of the subject matter and the complexity of the stakeholder interest in setting the right course for the UK, the recent Report was preceded by a separate offering published by the House of Lords Communications and Digital Committee on 6 March 2026, which sets out the House of Lords’ recommendations on the topic. By no coincidence, on 17 March 2026 (the day before the Report was published) the Chancellor issues a statement endorsing the UK as a future world leader in AI and quantum technology and announcing a £2.5 billion investment in the field.

Notwithstanding the clear depth of interest in the issues raised by Copyright and AI, or perhaps as an inevitable consequence of that interest, the Report paints a picture of caution, deferring responses on key issues in preference for further evidence gathering, before any significant legislative changes.

Key takeaways:

• No Immediate Legal Reform: The government has decided not to introduce immediate reform to UK copyright law regarding AI training; opting instead to "take the time needed to get this right". This approach reflects a lack of consensus among stakeholders and significant gaps in the evidence base, particularly regarding the economic impact of potential reforms. The government instead commits to further evidence-gathering and consideration of the developing impact of AI across the UK economy.
• Rejection of TDM Exception: The government has opted to move away from its earlier preferred proposal for a broad TDM exception under copyright law to allow for the training of AI models, with an opt-out for rightsholders. Most consultation responses - particularly from the creative industries - strongly opposed this, citing concerns that a broader exception would allow generative AI to learn from their works, without compensation, and in direct competition to them. Technology sector responses were also mixed, with some arguing the opt-out would be impractical or insufficiently competitive internationally. No alternative specific approach is proposed, with the government admitting to there being "no consensus on how these objectives should be achieved".   
• Removal of Protection for Computer-Generated Works: One of the only areas where the government has proposed legislative reform is in the context of computer-generated works ("CGWs"). Introduced in 1988, s.9(3) of the CDPA 1988 provides a reduced term of copyright protection for works generated solely by computers (including AI) without human input. In such circumstances, the ‘author’ is determined to be the person who made the arrangement necessary for the creation of the work. Feedback from the consultation was that copyright should incentivise and reward human creativity, not purely machine-generated output and accordingly it is proposed that protection should be removed, albeit that the government also proposed that it continue to monitor the impact of that protection and no timetable was set for the repeal of s.9(3) CDPA 1988. Happily, no change is proposed to the current protection afforded to works created with the assistance of computers or AI
• Focus on Transparency: Instead of legislative change, the government commits in the Report to prioritise measures to improve transparency around the data used to train AI models and encourage the development of licensing markets. A mandatory disclosure regime will not be pursued, despite strong support from the creative industries on this approach. There is support for labelling of AI-generated content, and adoption of technical tools and standards to help rightsholders control and license their works, but the government highlights that its approach must "promote clarity and enforcement for right holders, without disproportionate effects on AI development or deployment".
• Attention on Deepfakes: Given the growing risks to individuals posed by non-consensual "digital replicas" (i.e. deepfakes), the government will explore new protections to guard against deepfakes, considering the case for greater commercial protections, and whether these should form part of any wider safeguards or rights to personality within the UK. The government will launch a consultation in the summer to explore options for addressing the risks associated with deepfakes.

The Report reveals the government’s current position to be one of caution; a continuation of the "wait and see" approach observed over recent years, and perhaps an indication that as the AI landscape becomes increasingly more complex, so too are the interests of the various stakeholders. In a world where even those who might traditionally have found themselves at home on the ‘creatives’ side of the debate, now also find themselves increasingly adopting AI in the very process of creation. Perhaps, in those circumstances, the government’s commitment to further evidence-gathering; to continued consultation including monitoring of international developments might be more readily understood. How that sits with the ambition to make the UK a future world leader in AI and quantum technology remains to be seen.