Rethinking documentary heritage digitisation: an AI-centred economic model

The paper proposes a strategic shift in how we approach documentary heritage digitisation, framing it as an economic imperative driven by AI technology's growing need for high-quality training data. Current AI language models, whilst powerful, rely heavily on web-scraped content that represents only a fraction of human knowledge. This limitation presents an opportunity to reimagine heritage digitisation not merely as preservation but as a strategic investment in our digital future. Through AI-powered tools and workflows, digitisation can become more cost-effective and scalable, yet the humanities sector needs to fundamentally restructure its approach. The paper critically examines current academic practices that favour expensive, grant-funded flagship projects over sustainable, replicable workflows. To achieve meaningful scale in heritage digitisation, we need practical reforms in academia, including revised tenure criteria, micro-grant programmes, and stronger data-sharing practices. By learning from the life sciences and computer science sectors, humanities departments can develop more efficient approaches that balance scholarly rigour with practical scalability. This new economic model for digitisation promises to unlock the vast potential of our documentary heritage whilst fostering innovation and reducing technological monopolies in the AI sector.