Introduction
The UK Government has launched an open consultation to address the intersection of copyright law and artificial intelligence (AI), particularly focusing on a proposed exception for text and data mining (TDM) for AI training used commercially. This initiative seeks to clarify legal applications, balance the interests of creators and AI developers, and explore potential changes to existing copyright provisions.
Description
The UK Government has initiated an open consultation on copyright and artificial intelligence (AI), focusing on a proposed exception for text and data mining (TDM) specifically for AI training used for commercial purposes. This initiative aims to clarify the application of copyright law in the context of AI training [4], addressing concerns from creators about controlling the use of their content and seeking better remuneration [4], while alleviating the legal uncertainties faced by AI developers [4]. The consultation period remains open until February 25, 2025 [9], and includes 47 questions for stakeholders to address [9], covering practical aspects of opt-outs [9], technical standardization [9], and various related topics [9].
In addition to the proposed TDM exception, the government is revisiting the “computer-generated works” provision under the Copyright [2], Designs and Patents Act 1988 [2], which protects original works without a human author [2]. This provision has faced criticism for its contradictions [2], prompting the government to seek feedback from stakeholders [2]. The potential removal of this provision is under consideration [2], contingent on the consultation demonstrating sufficient benefits [2].
The proposed TDM exception seeks to align UK law more closely with the EU Article 4 of the Digital Single Market (DSM) Directive, enhancing access for AI developers while safeguarding the rights of copyright holders. This initiative emphasizes the importance of transparency regarding the sources of training material [4] [7], as AI developers may be required to disclose specific information about the datasets used [6], including details about the sources of content and materials ingested [6], and to document their text and data mining activities, providing clear and auditable records of copyrighted works accessed during AI training [6]. However, there is opposition to the introduction of the TDM exception [1], with concerns raised about the potential for mass scraping of copyrighted works without permission, which threatens the livelihoods of creators [1].
The government is also proposing an extension of the TDM exception in UK law, which currently allows non-commercial research to perform TDM activities without infringing copyright [8]. The proposed extension would permit commercial use [8], aligning with the EU exception under the DSM Directive [8]. This change aims to facilitate AI developers’ access to datasets for machine learning while balancing the rights of copyright holders [8], who would retain control unless they expressly reserved their rights [8]. Rights holders must be able to verify that their opt-out requests have been respected [6], which includes confirmation that AI developers have adhered to opt-out instructions in machine-readable formats or other agreed mechanisms [6].
The TDM exception allows rights holders to maintain control over their copyrighted content while facilitating the development of AI models in the UK [7]. If a rights holder has reserved their rights [7], a license will be necessary for data mining activities [7]. The initiative promotes good licensing practices to ensure that rights holders receive fair remuneration for their work while providing AI developers with broad access to high-quality datasets [3]. Transparency regarding the sources of training material is emphasized to foster compliance and trust between the two parties. The UK may implement a more stringent transparency framework than the EU [6], potentially requiring more detailed disclosure from tech companies [6].
Key proposals include enhancing transparency between AI developers and rights holders [10], allowing rights holders greater control over the use of their works for AI training [10], and ensuring AI developers have access to quality materials for model training [10]. The consultation suggests introducing a copyright exception for AI training for commercial purposes [5], which would provide legal clarity for AI developers regarding permissible material usage [5]. Alternative approaches to TDM have been considered [7], including the absence of an exception [7], which would require express licenses for AI developers [7], potentially hindering the UK’s competitiveness in the global AI landscape [7]. Conversely [7], unrestricted commercial TDM could enhance access to training materials but would limit copyright owners’ control and remuneration [7]. A “fair use” standard [7], similar to that in the US [1], could lead to costly litigation and significant changes to the UK copyright framework [7], potentially impacting the creative and media sectors [7].
The proposed TDM exception aligns with the EU general purpose TDM exception [7], while the UK’s existing exception for research purposes shares similarities with the EU research exception but has notable differences [7], such as its applicability to individual researchers [7]. The government is committed to engaging with industry and international standards to promote a collaborative approach and is seeking input on these distinctions.
To support compliance with standards [7], the government is considering the development of new technical tools to enhance transparency and may explore regulatory measures [7]. There is a focus on licensing as a means to ensure rights holder remuneration and facilitate access to quality training material for AI developers [7]. The government encourages collaboration between AI companies and creative industries to establish effective licensing practices [7].
Emerging issues include the generation of outputs during “inference,” where user inputs or live data interact with copyrighted works [6], and the use of synthetic data to train AI systems [6]. The application of copyright law to these scenarios may need clarification [6]. Potential new structures for data aggregation and licensing for AI training are being explored [7], including collective licensing and brokering services to streamline access for developers while ensuring rights holders receive fair remuneration [7]. The government also advocates for clear labeling of AI outputs [7], recognizing the technical challenges involved [7], and is seeking input on how to support the development of standards in this area [7].
The consultation invites stakeholders to provide feedback on a wide range of copyright issues arising from advancements in AI technology [7], reflecting the need for updated considerations in light of rapid developments in generative AI [7]. Stakeholders are encouraged to share their perspectives and evidence regarding the economic implications of the proposed changes [10], alongside broader engagement activities to capture a wide range of views [10]. The government recognizes the need for a balanced approach that protects creators while facilitating AI innovation [5], addressing concerns related to personality rights in the context of digital replicas [5], such as deepfakes [5], and evaluating the adequacy of existing legal frameworks [5].
Conclusion
The UK Government’s consultation on copyright and AI represents a significant step towards modernizing legal frameworks to accommodate technological advancements. By proposing a TDM exception and revisiting existing provisions, the initiative aims to balance the interests of creators and AI developers, ensuring fair remuneration and access to quality datasets. The outcome of this consultation could have profound implications for the UK’s position in the global AI landscape, potentially influencing copyright law, industry practices, and international collaborations [2] [12].
References
[1] https://www.britishcopyright.org/uk-government-launches-copyright-artificiai-intelligence-consultation/
[2] https://viewpoints.reedsmith.com/post/102jrv3/uk-government-launches-consultation-on-ai-and-copyright
[3] [https://www.gov.uk/government/consultations/copyright-and-artificial-intelligence](https://www.gov.uk/government/consultations/copyright-and-artificial-intelligence)/copyright-and-artificial-intelligence
[4] https://www.aoshearman.com/en/insights/ao-shearman-on-tech/ai-and-copyright-uk-launches-broad-consultation-on-potential-changes
[5] https://www.gov.uk/government/news/uk-consults-on-proposals-to-give-creative-industries-and-ai-developers-clarity-over-copyright-laws
[6] https://www.lewissilkin.com/insights/2024/12/17/uk-government-consultation-copyright-ai-and-the-proposed-opt-out-model-102jrlu
[7] https://www.jdsupra.com/legalnews/ai-and-copyright-u-k-launches-broad-5032931/
[8] https://mpaonline.org.uk/news/government-proposes-copyright-exception-for-ai/
[9] https://www.twobirds.com/en/insights/2024/uk/the-uks-new-consultation-on-ai-and-copyright-purrsuing-balance
[10] https://www.gov.uk/government/consultations/copyright-and-artificial-intelligence
