The MAIA COST Action Training School, “Creative, Critical and Practical Uses of Artificial Intelligence in Archaeology: the MAIA Training School at York” occurred September 2-4, 2025, King’s Manor, Department of Archaeology, University of York, UK. The participants underwent three days of intensive training and discussion regarding the future of Artificial Intelligence in archaeology.
Our first day was in the Digital Archaeology and Heritage Lab, where we had a morning workshop led by Dr James Stuart Taylor examining the ethical and political challenges of AI in archaeology, with regard to generative AI, chatbots, aerial detection, reconstruction of archaeological artefacts and more. Workshop participants engaged in intensive discussion over issues of ownership, permissions, FAIR/CARE principles, and legacy data.
In the afternoon Dr Peter Schauer gave a lecture on LLMs in archaeology with a practical exercise that drew from his work on large C-14 databases. Following these two sessions we had an intensive discussion regarding previous lectures, with a particular focus on ethical dimensions of the use of AI in archaeology. There was a reception held in the University of York Huntingdon Room, wherein participants, teachers, and members of the Archaeology Data Service and Department of Archaeology staff chatted.
The second day featured lectures from Miriana Somenzi on “The AI Scavenger: Digging Through Data to build Digital Twins” and Aida Himmiche on “The intelligent twin: Navigating the past, exploring the future with AI.” These sessions were highly technical providing some background knowledge about Knowledge Graphs, Ontologies, Digital Twins and how to navigate and use or think of using these technologies in the best way for Archaeology. Group discussions focussed on providing teaching materials to archaeologists to enable more engagement with the creation and use of AI tools for archaeological data.
The third day featured Dr Guy Schofield from the University of York in demonstrating Creative uses of AI for visualisation in archaeology. Guy walked us through specific workflows for creating museum interactives and visualisations, then we captured short clips of film on our phones and used AI tools to stitch together the clips to make a short documentary.
The final session was led by Dr Colleen Morgan in a “Blue Sky” exercise, wherein participants engaged in prefigurative imagining of optimal outcomes for AI in archaeology. Participants discussed the prompt, “Imagine a world ten years from now where the “problems” of AI and archaeology are completely solved” through the lenses of community, and were asked, “What does this ideal world look like? How is research conducted? How do people treat each other? How is the climate protected? How are problems solved?” As a group they determined the practical steps they can engage in to make their vision a reality.
We received good feedback from the training school participants and we look forward to future opportunities to learn more about AI and archaeology together.