Navigating the UK’s Evolving AI and Copyright Landscape
The intersection of Artificial Intelligence (AI) and copyright law is rapidly becoming a critical legal battleground, and recent developments in the UK offer a glimpse into the challenges and potential future direction of this complex field. Both the UK and US governments recognize the need to balance the burgeoning AI sector with the established rights of copyright holders, but are approaching the issue with distinct legal philosophies.
Earlier this year, the UK Government initiated a consultation on its proposed AI legal framework, specifically addressing copyright implications. Concurrently, the US Copyright Office released a report suggesting that utilizing copyrighted material for AI model training could constitute copyright infringement, specifically regarding the right to reproduce works. However, a key difference emerged: while the UK appears poised for a stricter stance, US AI developers can currently leverage the “fair use” doctrine as a defense, allowing continued use of copyrighted materials without explicit authorization.
A recent court judgment further clarifies the UK’s position. The case suggests that AI developers could be liable for secondary copyright infringement simply by making their models available within the UK, if it can be demonstrated that the model reproduces or stores copyright-protected images. Crucially, the court affirmed that “copying” under the Copyright, Designs and Patents Act (CDPA) can occur through intangible means, such as cloud storage, and that an intangible object can be considered an “article” for the purposes of the Act.
Though, it’s critically important to note the limitations of this ruling. The case was highly fact-specific, and several of Getty Images’ initial claims – including those relating to the training of the AI model itself – were withdrawn. A meaningful aspect of the case revolved around the location of the AI model’s training; as Stable Diffusion was trained outside the UK,establishing infringement within UK jurisdiction proved difficult. Consequently, the judgment doesn’t definitively address the legality of “scraping” copyrighted images for AI training, a central question in the debate.
The legal landscape remains fluid. Further litigation, potentially through appeals or new claims from Getty Images or other rights holders, is anticipated. For now, the outcome represents a partial victory for AI developers, but rights holders are actively seeking greater clarity on the UK’s approach.
Looking ahead,the UK Government is expected to publish a comprehensive report on copyright works in AI growth by March 2026.This report is anticipated to provide crucial guidance. One potential path forward, mirroring proposals from the EU, is the introduction of an exception to current copyright rules for text and data mining specifically for AI training.This would allow the use of copyrighted materials unless rights owners explicitly opt-out. While this proposal has faced criticism from creative industries,the current copyright framework,largely unchanged as the CDPA of 1988,is increasingly seen as inadequate in the face of rapid technological advancements. The recent judgment, coupled with ongoing government engagement, underscores the urgent need to reassess and modernize the UK’s copyright laws to reflect the realities of the AI era.
