WP6: AI for cultural heritage

This work package is about AI and our cultural heritage: the music, pictures, texts and recordings kept in archives, libraries and museums. Can AI help us find things in these huge collections and show them to people? We pay special attention to music and to the cultures of minorities, and we make sure AI is used with care.

This work package investigates how AI can help discover, organise and present the richness of cultural heritage collections. It develops intelligent tools for curating vast archives, with a particular focus on music and on the cultural expressions of minorities, while making sure that AI is used ethically, inclusively and responsibly in the heritage sector.

This work package investigates how AI can help discover, organise, and showcase the richness of cultural heritage collections. We develop intelligent tools for curating vast archives—with a particular focus on music and minority cultural expressions—while ensuring ethical, inclusive, and responsible use of AI in the cultural heritage sector.

Description

Archives, libraries and museums keep our shared past. The National Library of Norway has a very large digital collection and a strong lab for working with it. We want to connect it, and many smaller Norwegian collections, into one digital whole, and link that to the rest of Europe.

That needs smart tools that can sort and describe enormous collections on their own. One focus is music, where we combine machine learning with what musicologists know about how music works. Another is minority cultures: Norwegian folk music, Sámi joik and singing traditions from around the world.

Public institutions should lead the way here, so that these powerful tools are used fairly. Then institutions, artists and the public can explore how the past shapes creativity today, while giving credit where it is due.

Archives, libraries and museums are the institutions that preserve, share and communicate human culture. The National Library of Norway stands out with its large collection of high-quality digital material and its advanced Digital Humanities Lab. These resources, and many local collections across Norway, need to be brought into one coherent digital framework. That framework should connect to the European Cultural Heritage Cloud.

This requires automated, intelligent curation of very large collections. One focus is music, building on computational musicology at the University of Oslo in collaboration with the National Library. The models are hybrids: machine learning combined with musicology and with symbolic AI grounded in how people hear music. Particular attention goes to minority cultural expressions, including Norwegian folk music, Sámi joik and chant traditions from around the world. The aim is models for creating, curating and analysing catalogues that transfer to other heritage areas.

The public sector must lead in these technologies, so that powerful new capabilities are used ethically and inclusively. The approach gives institutions, artists and the public the means to draw on the archives. It lets them explore how past cultural expression influences creativity today, with proper citation and without plagiarism.

Archives, libraries, and museums (the “ABM sector” in Norway) are central institutions for preserving, sharing, and communicating human culture. The National Library of Norway is distinguished by its extensive collection of high-quality digital resources and the advanced Digital Humanities Lab. It is essential to integrate all these resources and numerous local collections in Norway into a cohesive digital framework and further connect them to the European Cultural Heritage Cloud.

This necessitates automated, intelligent curation of the vast collections. One focus will be on music, leveraging the cutting-edge computational musicology research at UiO in collaboration with the National Library. We will pioneer hybrid AI models by blending machine learning with musicology and cognition-based symbolic AI.

Particular attention will be dedicated to minority cultural expressions, including Norwegian folk music, Sámi joik and world chant cultures. We aim to create models for creating, curating, and analysing comprehensive catalogues transferable to other contexts and cultural heritage areas.

The public sector must lead in these technologies, ensuring that powerful new capabilities are used ethically and inclusively. This approach seeks to empower institutions, artists, and the public to leverage archives, exploring the influence of past cultural expressions on contemporary creativity while promoting respectful citation and avoiding plagiarism.

Research Questions

How can AI get much better at finding, sorting and showing what our collections hold?
How can AI protect and support the cultures of minorities, and what does AI do to cultural practice in general?
How can national institutions use AI research to guard against careless or exploitative uses of AI on heritage?
How can we link works together so that rights are clear and the influence of the past on today’s art becomes visible?

RQ6.1: How can AI be advanced to discover, organise and present the richness of cultural heritage and creative works, in every form, far more effectively than today?
RQ6.2: How can AI be used ethically and inclusively to protect and nurture minority cultural expressions, while taking AI’s wider effect on cultural practice into account?
RQ6.3: How can AI research help national institutions guard against unchecked and exploitative uses of AI on cultural heritage?
RQ6.4: How can systems that link cultural heritage and creative works be built so that rights are better managed and the influence of heritage on creativity today is made visible?

RQ6.1: How can we significantly advance AI to effectively discover, organise, and showcase the vast richness of cultural heritage and creative works in all forms?
RQ6.2: How can AI technologies be used ethically and inclusively to protect and nurture minority cultural expressions while considering AI’s broader impact on cultural practices?
RQ6.3: How can we advance AI research to enable national institutions to effectively guard against the risks of unchecked and exploitative AI use in cultural heritage?
RQ6.4: How can we develop interlinking systems within cultural heritage and creative works to enhance rights management and highlight their profound impact on contemporary creativity?

Approaches

CREATE: new AI tools that can transcribe, sort, connect and show huge collections of sound, images and text.
EXPLORE: how well today’s tools and our own tools preserve Norwegian culture, especially minority cultures.
REFLECT: on the risks AI brings to heritage, and write the rules for using it well.

CREATE: AI technologies that automatically transcribe, analyse, classify, interlink and present large unstructured multimedia collections, bringing out the richness of the works.
EXPLORE: how well existing and purpose-built AI systems preserve and nurture Norwegian cultural expressions, with a focus on minority cultures.
REFLECT: on the ethical and legal threats of AI in cultural heritage, and develop policies and guidelines for responsible, ethical and inclusive use.

CREATE: innovative AI technologies to automatically transcribe, analyse, classify, interconnect, and showcase vast, unstructured multimedia datasets, highlighting the richness of creative works.
EXPLORE: the effectiveness of current and custom AI systems in preserving and nurturing Norwegian cultural expressions, focusing on minority cultures.
REFLECT: on the ethical and legal threats of AI usage in cultural heritage and develop policies and guidelines to ensure responsive, ethical and inclusive use of AI.

Sustainability

This work helps with two of the United Nations’ goals for a better world. Tools that open up archives and collections make our heritage last longer and reach more people (SDG 11), and using them fairly, also for minority cultures, is part of building institutions people can trust (SDG 16).

The centre’s other commitments on sustainability, openness and reuse are on the open research page.

This work package contributes to two of the UN Sustainable Development Goals. Multimodal retrieval that opens archives and collections makes cultural heritage more durable and more usable (SDG 11), and doing so within ethical and legal limits, including for minority expressions, is part of building trustworthy institutions (SDG 16).

The centre’s other commitments on sustainability, openness and reuse are on the open research page.