Quantifying Culture: A Study of AI and Cultural Heritage Collections
the conceptual and practical interface between artificial intelligence and core archival imperatives
Description
Quantifying Culture investigates the potential of AI to enrich the management and interpretation of cultural heritage collections across Sweden. The project examines current GLAM digitalisation practices and develops methodologies for AI/ML-enhanced metadata generation, with a focus on ethics, diversity, and compliance with FAIR and international data standards. The research bridges the gap between computational methods and humanistic inquiry, aiming to demonstrate how AI can foster more inclusive, transparent, and engaging models of heritage curation.
This article investigates the pervasive issue of bias within AI-driven cultural heritage collections, emphasizing how digital technologies both inherit and amplify existing societal and historical prejudices embedded in analogue records. It outlines the multifaceted nature of bias—ranging from data selection and annotation to algorithmic design and user interaction—demonstrating how each stage of the AI pipeline can introduce or perpetuate distortions in representation. Through a critical review of current scholarship and practical case studies, particularly in image classification, the article evaluates technical strategies such as data augmentation, adversarial debiasing, and monitoring plans for bias mitigation. The findings reveal that while methods like noise injection and colour jittering can balance datasets and improve model fairness, effective bias mitigation ultimately depends on interdisciplinary collaboration between heritage professionals, subject experts, and data scientists. The article concludes that addressing bias requires an ongoing, holistic approach, integrating both technical and humanistic perspectives from data collection to model deployment. This ensures more inclusive, accurate, and ethically sound representations of cultural heritage, supporting the sector’s goals of diversity and accessibility for future audiences.
@article{Foka1959611,author={Foka, Anna and Griffin, Gabriele and Ortiz Pablo, Dalia and Rajkowska, Paulina and Badri, Sushruth},institution={Uppsala University, Human-Computer Interaction},journal={AI & Society: Knowledge, Culture and Communication},title={Tracing the bias loop : AI, cultural heritage and bias-mitigating in practice},doi={10.1007/s00146-025-02349-z},keywords={Cultural heritage, Artificial intelligence, Bias mitigation, Machine learning, Data augmentation, Interdisciplinarycollaboration, Image classification},year={2025},}
AI and Image: : Critical Perspectives on the Application of Technology on Art and Cultural Heritage.
AI and Image illustrates the importance of critical perspectives in the study of AI and its application to image collections in the art and heritage sector. The authors’ approach is that such entanglements of image and AI are neither dystopian or utopian but may amplify, reduce or condense existing societal inequalities depending on how they may be implemented in relation to human expertise and sensibility in terms of diversity and inclusion. The Element further discusses regulations around the use of AI for such cultural datasets as they touch upon legalities, regulations and ethics. In the conclusion they emphasise the importance of the professional expert factor in the entanglements of AI and images and advocate for a continuous and renegotiating professional symbiosis between human and machines. This title is also available as Open Access on Cambridge Core.
@book{Foka1990260,author={Foka, Anna and von Bonsdorff, Jan},institution={Uppsala University, Department of Art History},title={AI and Image: : Critical Perspectives on the Application of Technology on Art and Cultural Heritage.},isbn={9781009505468},year={2025},}
2024
AI, Cultural Heritage, and Bias : Some Key Queries That Arise from the Use of GenAI
Our article AI, cultural heritage, and bias examines the challenges and potential solutions for using machine learning to interpret and classify human memory and cultural heritage artifacts. We argue that bias is inherent in cultural heritage collections (CHCs) and their digital versions and that AI pipelines may amplify this bias. We hypothesise that effective AI methods require vast, well-annotated datasets with structured metadata, which CHCs often lack due to diverse digitisation practices and limited interconnectivity. This paper discusses the definition of bias in CHCs and other datasets, exploring how it stems from training data and insufficient humanities expertise in generative platforms. We conclude that scholarship, guidelines, and policies on AI and CHCs should address bias as both inherent and augmented by AI technologies. We recommend implementing bias mitigation techniques throughout the process, from collection to curation, to support meaningful curation, embrace diversity, and cater to future heritage audiences.
@article{Foka1908930,author={Foka, Anna and Griffin, Gabriele},institution={Uppsala University, Centre for Gender Research},journal={Heritage},number={11},pages={6125--6136},title={AI, Cultural Heritage, and Bias : Some Key Queries That Arise from the Use of GenAI},volume={7},doi={10.3390/heritage7110287},keywords={Cultural Heritage, AI, Artificial Intelligence, Generative AI, GENAI, Gender, Bias, Human-in-the-loop},year={2024},}
AI and Swedish Heritage Organisations : challenges and opportunities
Gabriele Griffin , Elisabeth Wennerström , and Anna Foka
AI & Society: The Journal of Human-Centred Systems and Machine Intelligence, 2024
This article examines the challenges and opportunities that arise with artificial intelligence (AI) and machine learning (ML) methods and tools when implemented within cultural heritage institutions (CHIs), focusing on three selected Swedish case studies. The article centres on the perspectives of the CHI professionals who deliver that implementation. Its purpose is to elucidate how CHI professionals respond to the opportunities and challenges AI/ML provides. The three Swedish CHIs discussed here represent different organizational frameworks and have different types of collections, while sharing, to some extent, a similar position in terms of the use of AI/ML tools and methodologies. The overarching question of this article is what is the state of knowledge about AI/ML among Swedish CHI professionals, and what are the related issues? To answer this question, we draw on (1) semi-structured interviews with CHI professionals, (2) individual CHI website information, and (3) CHI-internal digitization protocols and digitalization strategies, to provide a nuanced analysis of both professional and organisational processes concerning the implementation of AI/ML methods and tools. Our study indicates that AI/ML implementation is in many ways at the very early stages of implementation in Swedish CHIs. The CHI professionals are affected in their AI/ML engagement by four key issues that emerged in the interviews: their institutional and professional knowledge regarding AI/ML; the specificities of their collections and associated digitization and digitalization issues; issues around personnel; and issues around AI/ML resources. The article suggests that a national CHI strategy for AI/ML might be helpful as would be knowledge-, expertise-, and potentially personnel- and resource-sharing to move beyond the constraints that the CHIs face in implementing AI/ML.
@article{Griffin1758214,author={Griffin, Gabriele and Wennerstr{\"o}m, Elisabeth and Foka, Anna},institution={Uppsala University, Centre for Digital Humanities},journal={AI & Society: The Journal of Human-Centred Systems and Machine Intelligence},number={5},pages={2359--2372},title={AI and Swedish Heritage Organisations : challenges and opportunities},volume={39},doi={10.1007/s00146-023-01689-y},keywords={AI/ML implementation, Cultural heritage professionals, Cultural heritage management, Digital management of collections, Organization},year={2024},}
2023
Critically assessing AI/ML for cultural heritage : potentials and challenges
Anna Foka , Lina Eklund , Anders Sundnes Løvlie , and 1 more author
In Handbook of Critical Studies of Artificial Intelligence : , 2023
This chapter provides a critical examination of the promise of AI technology with a focus on museums and cultural heritage organisations. We argue that while AI shows great potential for digitalisation, collections management and curation, its implementation is a complex endeavour. First, we discuss artificial intelligence and machine learning technologies with great potential such as computer vision and natural language processing, as well as the implementation of AI for heritage encounters. We then identify a number of challenges in implementing these technologies—namely using technology to address the diversity of human memory and culture that is inherent in cultural heritage collections, but also issues of accessibility and technical know-how. Finally, we envision the future potential of AI for the digitalisation of heritage.
@incollection{Foka1825412,author={Foka, Anna and Eklund, Lina and Sundnes L{\o}vlie, Anders and Griffin, Gabriele},booktitle={Handbook of Critical Studies of Artificial Intelligence : },institution={IT University of Copenhagen, Copenhagen, Denmark},pages={815--825},title={Critically assessing AI/ML for cultural heritage : potentials and challenges},series={Sociology, Social Policy and Education 2023},doi={10.4337/9781803928562.00082},keywords={Artificial intelligence, Machine learning, Heritage, Curation, Diversity, Memory and culture},isbn={9781803928562},year={2023},}
2022
Women’s Metadata, Semantic Web, Ontologies and AI : Potentials in Critically Enriching Carl Sahlin’s Industrial History Collection
Anna Foka , Jenny Attemark , and Fredrik Wahlberg
In Emerging Technologies, Museums : Mediating Difficult Heritage , 2022
This chapter moves beyond claims of digital technology as a means for democratisation of knowledge and focus on archival online repositories of women’s history, concentrating on a Swedish case study: the collection of industry leader Carl Sahlin (1861–1943) at the Swedish National Museum of Science and Technology. The chapter contributes a detailed methodology for collection enrichment, including the possibilities and pitfalls of using emerging technologies, specifically AI, for classification and enrichment so as to open up new critical questions about historical women.
@incollection{Foka1643134,author={Foka, Anna and Attemark, Jenny and Wahlberg, Fredrik},booktitle={Emerging Technologies, Museums : Mediating Difficult Heritage},institution={Tekniska Museet },pages={65--88},title={Women’s Metadata, Semantic Web, Ontologies and AI : Potentials in Critically Enriching Carl Sahlin’s Industrial History Collection},doi={10.3167/978180073374900},keywords={Digital Cultural Heritage; Artificial Intelligence; Data and Information Science; Museum and Heritage Studies; Digital Humanities},isbn={978-1-80073-375-6},year={2022},}
Digital spetskompetens 2035 : Framtidsanalys för kompetensförsörjningen av digital spetskompetens
Fredrik Heintz , Jan Gulliksen , Amy Loufti , and 1 more author
2022
This is a governement report on digital expertise in higher education
Tillväxtverket och Universitetskanslersämbetet har fått i uppdrag från regeringen att tillsammans analysera och föreslå hur kompetensförsörjningen av digital spetskompetens ska kunna utvecklas både kort- och långsiktigt. Baserat på de tidigare rapporter som tagits fram inom uppdraget har vi fått i uppdrag att utveckla framtidsscenarier om hur tillgången till digital spetskompetens skulle kunna se ut samt komma med slutsatser och rekommendationer för framtida utvecklingen. Vårt tillvägagångssätt har varit att genomföra följande olika insatser: en SWOT-analys över nuläget, för att ha en startpunkt samt identifiera flaskhalsar och hinder för att realisera visionerna för digital spetskompetens, en trendanalys, som analyserar vilka stora trender med relevans för digital spetskompetens vi ser nu och inom de närmsta 10–15 åren baserat på vår egen omfattande forskning på området, en analys över viktiga policy-beslut som tagits och som skulle kunna tas, vilket visar på möjliga vägar framåt, inkluderande de tidigare rapporterna som genomförts i projektet och de internationella utblickar som gjorts inom uppdraget kring digital spetskompetens redan och analysera vilka förslag som skulle kunna vara genomförbara i en svensk kontext, samt ta i beaktande den input som vi fått från vår externa expertgrupp. Baserat på detta har vi låtit ta fram åtta framtidsscenarier på hur den framtida tillgången på digital spetskompetens kan se ut beroende på vilka policybeslut som tas samt hur trenderna utvecklar sig.
@techreport{Heintz1694572,author={Heintz, Fredrik and Gulliksen, Jan and Loufti, Amy and Foka, Anna},institution={Uppsala University, Department of ALM},note={This is a governement report on digital expertise in higher education },pages={55},title={Digital spetskompetens 2035 : Framtidsanalys för kompetensförsörjningen av digital spetskompetens},series={Rapport 0412},number={2022:2},keywords={digital spetskompetens, transformativa teknologier, AI},isbn={978-91-89255-99-9},year={2022},}
2020
AI for Digitalisation of Cultural Heritage : potentials and ethical challenges
Anna Foka , Lina Eklund , Anna Lowe , and 2 more authors
@inproceedings{Foka1501252,author={Foka, Anna and Eklund, Lina and Lowe, Anna and Skinner, Molly and Sundnes L{\o}vlie, Anders},booktitle={<em></em> : <em></em>},institution={Uppsala university},title={AI for Digitalisation of Cultural Heritage : potentials and ethical challenges},year={2020},}