User insights shaping machine learning applied to archives

Archives hold vast amounts of historical and cultural information, but navigating and extracting knowledge can be a daunting task. Machine learning (ML) offers immense potential to unlock these archives, yet its effectiveness hinges on understanding user needs. This paper explores how user insights can shape the development and application of ML in archives. Here “user” refers to editors and publishers who are crucial part of archival sorting and publication in the company. This paper emphasizes the importance of an iterative user centred design process to guide development and ensure user acceptance and empowerment. This approach reveals the distance between user expectations and functional integrity.