[Ohiodig] FW: [DHSI] New article on AI and sensitive colonial archives

[Carleton, Janet (she/her)]

May be of interest.
---
Janet M. Carleton | Interim, Libraries Communication & Marketing | Digital Initiatives Coordinator | Digital Initiatives | Mahn Center for Archives and Special Collections, Preservation & Digital Initiatives | OHIO University Libraries | Alden 333 | Athens, Ohio | 740.597.2527 | she/her | carleton at ohio.edu<mailto:carleton at ohio.edu> | Digital Archives<https://media.library.ohio.edu/> | Archives Finding Aids<https://archivesspace.ohio.edu/>

From: Institute <institute-bounces at lists.uvic.ca> On Behalf Of Lingjia Zhao
Sent: Tuesday, July 28, 2026 10:09 AM
To: institute at lists.uvic.ca
Subject: [External] [DHSI] 🆕 New article on AI and sensitive colonial archives

Use caution with links and attachments.
Dear all,
Apologies for cross-posting.
We are pleased to share our new open-access article:
👉 Should sensitive archival images be shown? The case for applying AI to colonial archives
🔗 https://doi.org/10.1007/s10502-026-09552-x
Published in Archival Science, the article explores how archives should manage colonial photographs depicting violence, nudity, coercion and other sensitive subjects.
Co-authored by Lise Jaillant, Matthew Kidd and Lingjia Zhao, the study draws on fifteen interviews with professionals working across GLAM institutions in the UK and continental Europe. It also includes a computer-vision test involving twenty colonial photographs from the Archives Nationales d’Outre-mer in France, allowing us to consider both the potential and the limitations of AI.
🔍 AI can help archives to:

  *   identify potentially sensitive material at scale
  *   improve metadata and description
  *   make difficult collections easier to discover
However, automated tools remain limited, particularly when the meaning of an image depends on historical and cultural context.
We therefore propose a practical framework combining:
🤖 automated triage
👤 human review and curatorial judgement
🏷️ clear labelling of machine-generated descriptions
⚠️ warnings, restrictions and takedown procedures
🤝 consultation and co-governance with source communities
The central argument is that AI can support responsible access, but archivists and source communities must retain authority over what is made visible, how it is described and when access restrictions are needed.
This research forms part of GENIAC: Generative Artificial Intelligence for Archival Images of the Colonial Period, funded by the British Academy. The project explores how AI can support responsible access to colonial images from the late nineteenth to the early twentieth century, working with source communities from formerly colonised countries.
📚 Citation:
Jaillant, L., Kidd, M. & Zhao, L. Should sensitive archival images be shown? The case for applying AI to colonial archives. Arch Sci 26, 36 (2026). https://doi.org/10.1007/s10502-026-09552-x
💬 We welcome perspectives from colleagues working with colonial collections, archival practice, digital humanities and responsible AI.
Best wishes,
GENIAC Research Team

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.library.ohio.gov/pipermail/ohiodig/attachments/20260728/f77dbf46/attachment-0001.htm>