Introduction
The rise of generative artificial intelligence (AI) tools has introduced significant challenges in copyright law, particularly concerning the authorship and use of AI-generated content. In the UK [3] [4] [5], the Copyright [1] [3] [4] [5] [6], Designs and Patents Act 1988 (CDPA) addresses these issues, but evolving interpretations and technological advancements necessitate a reevaluation of existing legal frameworks. The UK government has initiated a consultation to explore potential changes to copyright law [4], aiming to balance the interests of the AI sector and creative industries while ensuring legal clarity and promoting innovation.
Description
The emergence of generative artificial intelligence (AI) tools has raised significant copyright issues [3], particularly concerning the authorship of content generated by these systems [3]. In the UK [3] [4] [5], the Copyright [1] [3] [4] [5] [6], Designs and Patents Act 1988 (CDPA) includes a provision (section 9(3)) that attributes authorship of computer-generated works to the person who made the necessary arrangements for their creation [3]. This provision acknowledges that computer-generated works can have authors [3], but determining who qualifies as the author can be complex [3].
On December 17, 2024 [4], the UK government released a consultation paper addressing potential changes to copyright law in relation to AI [4]. This initiative aims to create a framework that balances the needs of the AI sector and creative industries [4], promoting human creativity and innovation while ensuring legal clarity for growth [4]. The consultation is open for public feedback until February 25, 2025 [4], and outlines four proposals: (0) maintaining the current copyright laws; (1) strengthening copyright by requiring licensing for all uses; (2) introducing a broad data mining exception allowing use of copyrighted works without permission; and (3) implementing a data mining exception with an opt-out mechanism for rights holders [4]. The government appears to favor option (3) [4], which would allow commercial data mining while giving rights holders the choice to restrict access to their works [4]. However, this proposal has faced significant opposition, particularly from musicians and content creators [5], who argue that the responsibility to opt out of making their content available for data scraping is impractical given the number of AI companies involved [5]. Critics express concern that such broad exceptions prioritize the growth of the AI industry over the copyright protections of creative professionals.
While the government’s objectives to protect rights holders while fostering innovation are welcomed, there is apprehension that the current copyright regime may hinder growth in both the creative and AI sectors [5]. The consultation emphasizes that compliance with copyright law by AI developers is essential for achieving the desired policy outcomes of control [6], access [1] [4] [6], and transparency [1] [4] [6]. Transparency is deemed essential for meaningful progress in policy and commercial negotiations [1], particularly regarding the use of copyright works in training generative AI systems [1]. New legal regulations are necessary to establish clear guidelines on how copyright material is utilized [1].
Existing UK copyright law is considered sufficient for licensing content to AI platforms [2], as there is no “fair use” provision [2], and “fair dealing” does not apply to commercial AI training [2]. Enforcement of current copyright law is advocated [2], alongside a call for increased transparency from AI developers regarding the materials used for training their models [6], including details on access and permissions from rightsholders [6]. The consultation paper highlights the need for transparency regarding how AI developers use copyrighted materials [4], yet it lacks clarity on what constitutes “minimum transparency standards.” It suggests that AI firms disclose specific works and datasets used [4], but acknowledges potential challenges in doing so [4].
Recent legal developments have questioned the application of section 9(3) CDPA [3], especially in light of evolving interpretations of originality in copyright law [3]. Traditionally [3], the ‘skill and labour’ test was used to assess originality [3], but this has shifted following the European Court of Justice’s ruling in Infopaq International A/S v Danske Dagblades Forening [3], which established a new standard defining originality as the author’s own intellectual creation [3]. This change has implications for works generated by AI [3], as the courts now require a demonstration of creative choices and personal expression to qualify for copyright protection [3].
The Court of Appeal’s decision in THJ Systems Ltd v Sheridan [3], while not directly addressing AI-generated works [3], provides insight into how authorship and copyright subsistence are evaluated [3]. The case involved software-generated charts [3], where the court determined that the originality of the work lay in the design choices made by the software developer [3], rather than the data populated by the user [3]. This suggests that similar principles could apply to AI-generated content [3], where the user’s creative input may be critical in establishing copyright [3].
To address the challenges posed by generative AI, a technical solution for expressing creators’ intent regarding the use of their works in AI training data sets is proposed through the embedding of metadata in image and video files [2]. This approach [2], alongside the Text and Data Mining Reservation Protocol [2], could serve as a substitute for formal licensing agreements [2], facilitating an opt-in method that is both technically feasible and scalable [2]. The government is urged to promote such licensing practices that would benefit both creators and AI developers [6], thereby supporting the creative and cultural industries [6]. Ultimately, the government believes that works exhibiting human creativity [3], even if generated by AI [3], could still be protected under the general originality standard without the need for a specific statutory rule [3]. However, there remains a risk that the UK may align more closely with the US’s light-touch regulatory stance if it fails to adequately address the concerns of the creative industries [5].
Concerns about bias and misrepresentation in AI systems are also highlighted [4], with examples of past failures in AI recruitment and image recognition due to skewed training data [4]. A broad data mining exception could potentially diversify the data used for training AI [4], but it may also undermine rights holders’ ability to receive compensation for their work [4]. The financial implications for creative industries could be significant [4], as the AI art market is projected to grow substantially [4]. Some platforms are emerging to help rights holders manage the use of their work in AI training [4], such as Spawning, which has developed a tool that allows artists to opt out of datasets used for training AI [4], addressing transparency issues and enabling artists to protect their rights [4]. Additionally, the establishment of an AI Copyright Hub is suggested to integrate licensed content with technical tools [1], creating a cohesive and internationally recognized framework for legal access to copyright material in generative AI development [1].
Conclusion
The consultation initiated by the UK government represents a critical step in addressing the complex interplay between copyright law and generative AI. The proposed changes aim to foster innovation while safeguarding the rights of creators. However, the potential for significant impacts on the creative industries, particularly concerning compensation and control over copyrighted works, necessitates careful consideration. The outcome of this consultation could shape the future of copyright law in the UK, influencing both the AI sector and creative professionals.
References
[1] https://www.alcs.co.uk/news/alcs-submission-to-governments-ai-and-copyright-consultation/
[2] https://iptc.org/news/iptc-and-plus-respond-to-uk-consultation-on-copyright-and-ai/
[3] https://www.jdsupra.com/legalnews/what-makes-an-ai-generated-work-original-4921359/
[4] https://itsartlaw.org/2025/03/03/remodelling-the-uks-gold-plated-copyright-regime-and-its-impacts-on-creative-industries-and-ai-training/
[5] https://www.verdict.co.uk/is-the-uks-ai-copyright-proposal-another-attempt-at-bridging-the-eu-us-ai-regulatory-chasm/
[6] https://www.dacs.org.uk/news-events/copyright-and-ai-consultation-dacs-responds
