Open Heritage in the Age of Artificial Intelligence

The Prodigious Value of Sharing Heritage

Cultural heritage institutions (CHIs) such as museums, libraries, and archives, carry the important public-interest mission of providing access to the heritage in their collections. Their mission is grounded in broad public access and extensive human engagement, the existence of a vibrant cultural commons, and heritage enjoyed as a shared resource. CHIs understand that, as a representation of who we are, heritage helps define our shared humanity and that providingaccess to heritage generates numerous benefitsfor society. As such, CHIs, as stewards of trustworthy information imbued within knowledge and culture, have a crucial role to play in growing and protecting the commons and enabling uses that benefit humanity. 

AI—the Major Disruptor 

The recent emergence of generative artificial intelligence (AI) has had a phenomenal impact on CHIs, especially as the content they steward is accessed and used for AI development. One key question centers on the conditions, norms, and expectations governing the use of openly licensed and public domain cultural heritage materials as data for AI development. Examples include the open datasets available atAI4Culture, Hugging Face’sBigLAM, Harvard’sInstitutional Books, Authors’ AllianceThe Public Interest Corpus, as well as, potentially in the future, the<a href="https://www.dataspace-culturalheritage.eu/en/news/feasibility-study-paves-the-way-for-a-european-books-data-commons” rel=”nofollow noopener” target=”_blank”>European Books Data Commons. This is promptingdiscussionsaround who can access and reuse those data, and for what purpose. Heritage collections often yield—by virtue of the stewardship duties that underpin them—diverse, high-quality, contextualized, and well-curated datasets. While inclusion of heritage collections data could expectedly yield better,more representativeAI systems, CHIs face many challenges, including: 

• Infrastructure challenges, which threaten the economic sustainability of the open infrastructure and open ecosystem, including:

• • Increased bot traffic by dominant, profit-driven AI companies putting a strain on the infrastructure and increasing knock-on costs (as reported inAre AI Bots Knocking Cultural Heritage Offline?, GLAM-E Lab);
  • No sharing back of any value generated (as discussed in “Safeguarding Cultural Rights and Redistributing the Value Generated by AI Use of Cultural Data” in the 2025 UNESCOReport of the Independent Expert Group on Artificial Intelligence and Culture).
• Commercial AI practices at odds with the values and norms of the commons, including:

• • Lack of transparency and attribution, (as tackled at theOulu Löylyevent’s working group,May the Source be with You), decontextualization (as noted inArtificial intelligence and creativity, a Report by the Special Rapporteur in the field of cultural rights, Alexandra Xanthaki), and loss of provenance trail, due to the opacity of AI models and systems; 
  • Lack of information and understanding about how the content is being reused, whether for training, inference, or search; or whether the content is being memorized by the model or retained post-training (as mentioned in the CC blog postFrom Signals to Infrastructure: Strengthening the Commons for the AI Era);
  • Further enclosures of the content into proprietary systems (as it happened foracademic research on X/Twitter).

On the surface, this points to a conflict between the practices of commercial AI developments and CHIs’ commitment to provide open access and commons-based expectations ofreciprocityand respect for the social contract underpinning sharing of heritage in the commons. It reveals how trust is slowly being eroded. And all this is happening against a backdrop of high legal uncertainty around copyright and, by the same token, the application of CC licenses and tools.

CHIs More Hesitant to Share in the Commons

We can map reactions to AI scraping along a spectrum from wide embrace to protective, defensive closing up. This is well reflected in Europeana’s reportMaking Public AI reality: How cultural heritage can lead the way, which identifies two groups at each end: boundary setters and opportunity seekers. The onset of AI traffic basically leaves CHIs with two options: 

• Allow unrestricted extraction by all AI developers, or
• Restrict access in blunt and overbroad ways (“defensive enclosure”).

In fact, some institutions, e.g., theNational Library of the Netherlands, and Trove, the National Library of Australia’s open access API (seehereandhere), have responded by closing up access, revealing a paradox: CHIs, whose mission is to share their collections, end up locking up heritage. This conundrum is rattling the fragile scaffolding ofopen heritage, as content that would have otherwise been made available openly is now being further restricted, with access to heritage in the public interest as collateral damage. Indeed, this tension is acknowledged in theOpen Heritage Statement. To recall, the concept of open heritage encourages access and reuse of heritage free ofunfair, unnecessarybarriers; it does recognize that openness is neither absolute nor monolithic. The Statement’s preamble paragraph (s) emphasizes ethical and responsible AI and states:  

[W]e are attentive to the challenges and opportunities brought about by emerging technologies, including artificial intelligence and future technologies, and the need for more ethical and responsible artificial intelligence with respect to access and reuse of public domain heritage.

This serves as a timely reminder that when CHIs embrace openness, it is as a means to an end—not open for open’s sake. CHIs often embrace openness in accordance with its underlying principles, which recognize nuance, contextuality, and the interplay among various governance frameworks. That said, closing up completely is not the answer (it’s inefficient at best and harms public interests at worst) and risks undoing 20+ years ofopen heritage using CC licenses and tools. We thus need to move beyond a reactive enclosure pattern towards a more sustainable sharing equilibrium. 

From Fragmentation to a Sustainable Commons 

We know that many CHIs want to continue sharing their collections while ensuring that AI developers use them responsibly. Against this backdrop, how can CHIs make their heritage collections openly and equitably accessible in a way that aligns with their mission as responsible stewards? How can CHIs “actively shape the conditions under which their collections are accessed and reused by AI systems, without compromising their commitment to openness” asOpen Future and Europeana state? More specifically, under what conditions should heritage data be made available for AI training and use? By whom? For what purpose?

As a response to these questions, CC is developing CC Signals, new tooling which will restore agency in CHIs as data stewards, while maintaining the open nature of heritage shared in the commons. CC Signals will attempt a balancing act between: 

• Preserving access to valuable heritage resources, and
• Ensuring that the institutions and communities who steward them remain active participants in shaping the AI ecosystem. 

CC Signals tooling will create a middle path between the current binary of “free for all vs defensive enclosure.” CC Signals will enable this by establishing standardized, prosocial conditions for access and reuse in an AI context. As part of this work, CC is exploring avenues for combining contract and copyright law to give CC Signals legal force. All forms of CC Signals will incorporate protections for public interest reuse.  Put simply, CC Signals would set and uphold the boundaries that the commons needs in order to thrive. CHIs are well positioned to test and implement CC Signals as part of ethical frameworks for sharing heritage in the commons for AI use. CHIs are key actors to help shape an AI ecosystem that fosters collaboration, ensures equitable access, and yields long-term public benefit. 

How You Can Help

While we don’t have all the answers yet, we know we need to start having ambitious conversations to make sure CHIs’ voices inform our increasingly AI-mediated world. Today, you can contribute to shaping the AI that works for open heritage: 

• Sign theOpen Heritage Statementto show your support for ethical and responsible AI.
• Learn more about theCC signals projectand/or how you can get more involved.  
• Help build and support the commons’infrastructuresand prosocial technologies that align with open values and embody the open spirit.
• Invest in the open infrastructure.