🔍 In this episode of Humans of AI, Henk van Stappen and Tobias Hendrickx take us inside MODAL, a project exploring how AI can make large born-digital cultural heritage archives more accessible and discoverable. They discuss working with smaller, open-source models, the challenges of speech-to-text for Flemish dialects, and why cultural heritage institutions don’t always need the biggest AI models.
📌 About Our Guests
🌐
https://advn.be/nl/over-advn/projecten/modal-project Henk Vanstappen is co-founder and senior consultant at Datable, where he advises cultural heritage organisations on digital strategies connecting technology, data and collections. With a background in Arts and Philosophy and Information Sciences, he has worked on digital heritage projects since 2004, with expertise in digitisation, metadata, digital collections and emerging technologies such as machine learning and generative AI.
Tobias Hendrickx works at MoMu – Fashion Museum Antwerp, where he is involved in the management and accessibility of the museum’s digital archives. Within MODAL, he explored how AI can help make born-digital fashion archives more discoverable, including through speech-to-text and metadata generation.
MODAL (Metadata and Access to Digital Archives using Large Language Models) is a two-year research project supported by the Flemish Government, investigating how generative AI can support the cultural heritage sector. The project aims to develop a practical application for archivists and collection managers, helping them explore large born-digital archives and make faster decisions about processing and eventually opening up their collections. Alongside the technology, MODAL also focuses on sharing knowledge and encouraging collaboration around AI and cultural heritage.
#AI #ArtificialIntelligence #GenerativeAI