European Journal of Cultural Management and Policy
Paula Westenberger and Despoina Farmaki
Abstract
Artificial intelligence (AI) is revolutionising our relationship with cultural heritage, enhancing access to, engagement with and preservation of collections and heritage sites. AI is also being used as a valuable research tool in the context of heritage collections. However, as materials protected by copyright may be used in AI development, training and use, copyright law can become an obstacle to important AI deployments in the heritage sector, an area which is currently understudied from the United Kingdom (UK) perspective. This article explores the intricate interplay between cultural heritage, AI and copyright law, demonstrating the main copyright law and policy challenges facing cultural heritage professionals and researchers in using AI in the UK for heritage research. It highlights the complexity and uncertainties as regards the current Text and Data Mining exception in the UK Copyright, Designs and Patents Act 1988 (UK CDPA), emphasising the need for an improved legal framework that balances copyright protection with the benefits of AI for cultural heritage research and management. It also reveals the underrepresentation of the heritage sector in AI regulation and copyright policy discussions in the UK. This exploration underscores the imperative for an inclusive policy dialogue that considers the perspectives and evidence of the cultural heritage sector in its full breadth and diversity (including related researchers) in shaping copyright law reform and AI regulation, and for further research to be carried out in this field.
Introduction
The adoption of new technologies in the cultural heritage sector often raises issues around copyright law, particularly whether current legislation is fit for purpose. Historically, copyright law has constantly adapted to address technological advances (Gervais et al., 2024, p. 28). With the increasing use of Artificial Intelligence (AI) across sectors, copyright laws are once again under scrutiny. However, there is a specific need to address the unique copyright challenges AI presents to the UK heritage sector, which are currently underexplored.1
Cultural heritage institutions, entrusted with the preservation of the legacies of humankind, are navigating complex legal terrain in their pursuit of modernisation. AI is revolutionising the forms of usage of cultural heritage materials, including as a tool for preservation, research, and dissemination. The volumes of data now available create traction to digital humanities approaches, facilitating the processing of data in scale “to yield new analytical insights that were not possible at the level of individual documents and sources” (Ahnert et al., 2023). While the importance of digital heritage collections as AI data2 is being discussed in current academic and sector-specific debates, there is scope for investigation of related copyright challenges in the UK.3 Since copyright materials may be used in AI development, training and use, the current issues and uncertainties in copyright law may pose obstacles and ultimately hinder important AI applications in the traditionally risk-averse heritage sector. Although these issues are being explored in Europe,4 the UK lags behind, making this a crucial and novel area of investigation in a jurisdiction that can influence legislations worldwide. This paper therefore aims to contribute to current underdeveloped UK policy discussions on AI and copyright with such a heritage perspective, by analysing the particular shortcomings of, and proposing solutions for, the UK Text and Data Mining (TDM) exception for non-commercial research, which we argue is not currently fit for purpose when applied specifically to heritage research.
Section Artificial intelligence and cultural heritage highlights the importance of AI in the heritage sector with UK examples. Section Copyright challenges arising from AI uses in the cultural heritage sector in the UK examines key issues in UK copyright law, particularly the TDM exception, crucial for AI development and knowledge discovery in heritage research. Section Current copyright and AI policy and regulation efforts in the UK: scope for further heritage sector participation addresses the state of current policy discussions on copyright and AI regulation, emphasising the need for greater involvement and evidence from the cultural heritage sector.
Artificial intelligence and cultural heritage
Although there is no generally accepted definition for AI (Sheikh et al., 2023; Guadamuz, 2024), for the purpose of this paper it refers to computer programmes that perform tasks associated with human intelligence, such as language comprehension, image recognition, and learning from experience (Pavis, 2023). AI is the umbrella term including machine learning, whereby a computer programme is taught to identify patterns in data and apply this knowledge to new data (Drexl et al., 2019; Iglesias Portela et al., 2019).
Read the article here:
