Impulse paper: Publishing cultural heritage data in the age of AI
This joint impulse paper by the Open Future Foundation and the Europeana Foundation explores how cultural heritage institutions can navigate data publishing in an era increasingly shaped by AI. Commissioned by the Europeana Foundation to the Open Future Foundation as part of, and a contribution to, the ongoing Alignment Assembly on Culture for AI—a collective intelligence and consultation process that has been taking place within the common European data space for cultural heritage since May 2025—it addresses one of the sector’s most pressing questions: under what conditions should heritage data be made available for AI training and use? One of the key topics that emerged from the Alignment Assembly concerns the opportunities and challenges of positioning heritage data as responsible AI training material. This Impulse Paper explores that topic further, focusing on generative AI and its implications for data sharing in the cultural heritage sector.
The paper is organized in two parts. The first outlines the relevant technological and legal context, examining how AI systems interact with cultural heritage data during both training and deployment phases, and analyzing the practical limitations of current copyright and technical mechanisms for managing automated access. The second part proposes a differentiated access model for cultural heritage data.
The Operational Challenge
Cultural heritage institutions face a fundamental shift in their operational context. Access models designed for human-scale use—discovery and consultation of individual works—now confront demand for collection-level access at an industrial scale. AI has created unprecedented demand for bulk access to digitized collections, yet institutions find themselves with limited tools to manage this transformation while staying true to their public mission.
The paper reveals a core tension: while there is broad support for making publicly funded heritage data as open as possible, privacy concerns, rights management, and the realities of large-scale automated reuse complicate this commitment. Instead of relying on legal or technical restrictions, the paper grounds its decision-making framework in a set of public-interest principles that reflect the mission of cultural heritage institutions. These principles emphasise maintaining open access to knowledge wherever possible, supporting innovation and research through responsible data sharing, and safeguarding the trustworthiness, authenticity, and authority of cultural heritage as a public information resource. The proposed framework uses these principles to help institutions evaluate when open access is appropriate, when conditional access may be justified, and how to manage large-scale AI reuse in ways that remain aligned with their public mandate.
A Framework for Action
Rather than relying solely on legal restrictions or technical barriers, the paper develops a framework to help institutions decide whether—and under what conditions—to make collection data available for AI training. This differentiated access model aims to balance commitment to open access with the need to manage new forms of large-scale reuse, while contributing to a sustainable information ecosystem. Such a framework would help institutions balance their commitment to open access and public information provision with the need to manage new forms of large-scale reuse that come with the rise of AI.
Invitation for Feedback
This framework is offered as an invitation for dialogue and collective exploration, responding to the participatory spirit of the Alignment Assembly. The model is a starting point for discussion, not a prescriptive blueprint. Its feasibility depends on collective assessment by institutions across the Europeana Initiative, the common European data space for cultural heritage, policy makers, and the broader sector. The Europeana Foundation and Open Future invite feedback to inform future work toward a shared framework for conditional access that aligns with cultural heritage institutions’ mission, values, and sustainability needs.
