Towards Machine-Assisted Sensitivity Review: Structuring Knowledge, Precedent and Expertise for Historical Records

This PhD will investigate how artificial intelligence (AI), Privacy Enhancing Technologies (PETs), knowledge management systems, similarity checking and other data-driven approaches can support more consistent, scalable and trusted sensitivity review of digital records selected for transfer to The National Archives.

The research will explore how precedents, expertise and contextual knowledge from previous review decisions can be identified, captured and reused to improve the consistent application of sensitivity criteria across organisations and over time.

Drawing on archival science, records management, computer science, information governance and public policy, it will address the technical, organisational and regulatory challenges of digital sensitivity review. Working with sensitivity reviewers across government, the project will also examine review practices, decision-making processes, sources of expertise and influencing factors. It will investigate how decisions, rationales, exemptions, risks and contextual information can be captured and reused as organisational knowledge. This will include through knowledge graphs and collaborative platforms, and how AI can identify relevant patterns and precedents.

The project will combine empirical research with practical prototyping, enabling the student to develop and evaluate proof-of-concept solutions with records management practitioners.

For more information about this project, please get in touch with The National Archives’ project team: Balint Csollei and Chris Royds

forms.cloud.microsoft

Submit your Expression of Interest here

Please submit your expression of interest using this form. The closing date for receiving EOIs is 17.00 on Friday 9th October 2026.