AI is becoming a dominant interface to knowledge speak with a clear, distinct and shared voice.If our and culture – yet the development of AI systems is in sector does not act now, the risks are manifold, and the hands of a few large private companies that are extend well beyond our sector alone.opaque and too often detached from public value.The time for action to address structural asymmetries For AI and Europe in the AI ecosystem is now.Now more than ever, AI is putting the social contract of trust at risk.
In sectors AI systems developed without the active such as education and research, the use of AI without contributions of the cultural heritage sector meaningful human involvement can erode the risk becoming powerful but fundamentally trusted relationships that sustain knowledge creation unreliable.While able to scale efficiently, it is at and exchange – such as those between teachers and the cost of trust, nuance and evidential integrity – students.
Cultural heritage institutions are credible precisely the qualities needed for AI to serve the within the communities they operate in and can public interest.Equally Europe risks reinforcing leverage their position and expertise to advance current commercial AI dynamics if it does more trusted, accountable and pluralistic AI.not champion a distinctly different approach grounded in the public interest.This concern There is also a sense of urgency to our joint action.
is increasingly echoed in Public AI scholarship As sectoral strategies take shape and the EU defines and European digital policy debates: without its next long-term budget, including significant explicit public-value governance, the investments investments in AI, the cultural heritage sector must in computation, data and deployment may end
45 up reproducing rather than reshaping current concentrations of power in the AI ecosystem. If AI systems continue to rely mainly on English, they will reflect an incomplete view of the world, sidelining Europe’s linguistic and cultural diversity. Edzésre váró gyerekek a BVSC-pályán by Hemző Károly (1928-2012) - 1960 - HNM-PCC Hungarian National Museum, Hungary - CC BY-NC-SA.
46 For the open knowledge ecosystem For cultural heritage institutions The open knowledge ecosystem is already Cultural heritage institutions continue to bear under pressure, and will be at risk without the costs of their data being extracted by a shared framework through which public opaque AI systems without consent, attribution institutions – from cultural heritage to media, or compensation.
This will put at further risk publishing, education, research and public the institutions’ ability to afford to provide a sector information – can align, organise and act robust public service and therefore jeopardise collectively.AI systems will continue to extract their relevance in the AI era.
As AI increasingly and repurpose content from public institutions determines what users see and engage with, the without consent, redistribution of benefits, visibility and diversity of cultural content will also compensation or attribution, shifting the costs be at risk.of infrastructure, preservation, governance and trust onto the institutions whose knowledge they exploit.
This puts at risk the willingness of cultural heritage institutions to keep their data open and the sustainability of the broader ecosystem of knowledge organisations, media and other information providers.
47 The cultural heritage sector must use its knowledge Coalition building and policy alignment and experience, practices and values to promote and develop Public AI – starting with the common As sectoral strategies develop – particularly European data space for cultural heritage.
The the upcoming AI Strategy for the Cultural Alignment Assembly has shown both the urgency and Creative Sectors to be announced by the for action and the sector’s wish for a distinctive European Commission in 2027, as outlined in the contribution, as well as the role the data space can Apply AI Strategy and the Culture Compass – and play in enabling it.
This is a more confident, political as the EU defines its next long-term budget with and public-facing role for the data space: as rights significant AI investment, this paper offers the broker, trust infrastructure, literacy engine, technical cultural heritage sector a clear, shared basis for partner and collective voice for cultural heritage.In joint advocacy.
It will support coalition building the months ahead, the data space will support the and policy alignment, not only within our sector cultural heritage sector in the following areas of but across the broader public knowledge and AI action: ecosystems.
48 Knowledge sharing and capacity building Advancing collective ownership of the Public AI vision The data space and the Europeana Initiative will mobilise its wide network – including the This paper is a starting point.
Its messages, and Europeana Network Association communities how to operationalise them through the data and the Europeana Aggregators’ Forum – to space, will be further developed with input develop thought leadership and strengthen from the data space community, the Europeana AI literacy across the sector.This will include Initiative and relevant partners, reflecting the developing protocols and frameworks – for collaborative spirit of the Alignment Assembly.
example for sharing cultural heritage data for The data space will continue to build collective AI use within a reciprocal logic – and guidelines, ownership of the Public AI vision in the heritage training and practical tools to contribute to a community, and we welcome ongoing reactions, coherent European approach.It will also focus reflections and proposals.
on areas for further investigation to mature the ideas put forward in this paper: from the The task ahead is clear: align around a shared environmental impacts of AI to exploring position, use the data space to build practical differentiated access models.capacity, and shape a Public AI ecosystem where cultural heritage professionals continue to deliver open access to knowledge while safeguarding