Artificial Intelligence and Cultural Heritage: A Review of Opportunities, Risks, and Ethical Challenges
TL;DR
Artificial intelligence should be understood not simply as a technological tool but as a socio-cultural intervention whose deployment requires collaboration among technologists, heritage professionals, researchers, governments, and source communities.
Abstract
Artificial intelligence (AI) is increasingly transforming the documentation, preservation, interpretation, and public engagement of cultural heritage. This review examines the opportunities, risks, and ethical challenges associated with the application of AI in cultural heritage management and research. Drawing on interdisciplinary literature spanning heritage studies, computer science, digital humanities, museum studies, and information ethics, the review synthesizes evidence on the use of machine learning, computer vision, natural language processing, generative AI, and three-dimensional technologies in heritage contexts. The findings indicate that AI can enhance the digitization and classification of cultural collections, reconstruct damaged or lost heritage assets, support archaeological analysis, improve multilingual access to heritage resources, and create more interactive experiences for museums and heritage sites. AI also offers opportunities for predictive conservation through automated monitoring of environmental conditions, material deterioration, and structural changes. However, these benefits are accompanied by significant risks, including algorithmic bias, inaccurate reconstruction, cultural misrepresentation, intellectual-property disputes, data privacy concerns, technological dependency, and the marginalization of communities whose cultural knowledge is incorporated into AI systems without meaningful participation or consent. Generative AI presents additional challenges because synthetic representations may blur distinctions between historically documented evidence and machine-generated interpretations. The review therefore emphasizes the importance of human oversight, transparent documentation of AI-generated outputs, culturally sensitive data governance, community participation, provenance tracking, and equitable access to digital heritage technologies. Overall, AI should be understood not simply as a technological tool but as a socio-cultural intervention whose deployment requires collaboration among technologists, heritage professionals, researchers, governments, and source communities. Responsible and inclusive AI governance can enable technological innovation while protecting the authenticity, diversity, ownership, and cultural significance of heritage resources.