Abstract
This talk recasts the future of contemporary artificial intelligence through the lens of a historiography enabled by a new history of AI. It explores how approaches to AI in the second half of the twentieth century turned on making sense of data from the past. It considers the implications of this view for the intersection between contemporary AI and history through three scenarios: collision, where contemporary AI becomes de facto the new discipline of history; parallel, where contemporary AI and history co-exist as separate and distinctive historical disciplines; and transformation, where contemporary AI and history take a mutually interdependent path where each gives shape to the historiographical concerns of the other. It argues for the transformative potential of mutual historiographical recognition as AI developers seek ways to work with incomplete even prejudicial data from the past more effectively, fairly, sustainably, and even securely. Working with smaller, more complex datasets will for example help AI companies to reduce their data centre footprints.
Bio
Marnie Hughes-Warrington AO is Bradley Distinguished Professor of History/Philosophy at Adelaide University. Since 2019 she has published a series of papers and books on the relationship between history and artificial intelligence. Her most recent books are Artificial Historians (with Lewis Yurlapurka O’Brien and Anne Martin, Routledge, 2025), which was shortlisted for the International Commission for the History and Theory of Historiography book prize, and History and Artificial Intelligence (Cambridge University Press, 2026). Her books have been downloaded over 105,000 times since 2022 and she has been associated with $18m AUD in grants. This talk offers a glimpse of her new book project.
