EMILiA: Making Email Archives Accessible through Intelligent Entity Recognition and Automated Anonymization

Speaker: Nico Beyer, Freie Universität Berlin / Archives of the Max Planck Society

Date: 17 February Time: 13:00–14:00 CET

Short introduction: This webinar introduces EMILiA, a toolbox that helps archives make historically and legally relevant email collections accessible with the help of AI. The webinar focuses on context-sensitive entity recognition, automated metadata enrichment, and semi-automated anonymisation that enable legally compliant access to email archives. It will also demonstrate how AI can support appraisal, analysis, and the long-term preservation of complex email records in practice.

Registration for the webinar is CLOSED. Video presentation now available (members only).

Back to top

 

eArchiving 3D models: an end-to-end preservation solution for Digital Cultural Heritage (DCH)

Speakers: Janet Anderson, Highbury R&D; Sven Schlarb, Austrian Institute of Technology (AIT); Stephen Mackey, Penwern

Date: 3 March Time: 13:00–14:00 CET

Short introduction:This webinar presents an end-to-end eArchiving solution for the long-term preservation of and access to 3D models as part of our Digital Cultural Heritage (DCH). It starts by covering the necessary groundwork to be carried out, leaning heavily on a key comprehensive report by Prisma colleagues https://digital-strategy.ec.europa.eu/en/library/report-long-term-preservation-3d-data-cultural-heritage. It then introduces recent work on standards, including the E-ARK Content Information Type Specifications (CITS) for 3D product models and cultural heritage objects, developed within the eArchiving Initiative in close collaboration with the DCH community. The webinar concludes with a practical demonstration showing how 3D models can be archived, managed, and rendered using the eArchiving Reference Implementation (earkweb) and other tools, illustrating a complete preservation workflow in practice.

Registration for the webinar is CLOSED. Video presentation now available (members only).

 

The use of AI and Machine Learning tools to enrich metadata and recreate life-courses at the Danish National Archives

Speakers: Jakob Leander Humlegaard & Tobias Kallehauge, Danish National Archives

Date: 9 April Time: 13:00–14:00 CET

Registration for the webinar is CLOSED. Video presentation now available (members only).

Back to top

 

Compute-to-Data: A New Paradigm for AI in Archives

Speaker: Victoria Lemieux, University of British Columbia

Date: 27 May Time: 15:00 CEST

Short introduction: Archives are under growing pressure to enable AI-driven discovery while safeguarding privacy, rights, and public trust. Current AI models, which rely on centralizing and extracting data, are fundamentally misaligned with archival principles and create unacceptable legal and ethical risks. This presentation introduces an alternative paradigm: bringing AI to the archive, where computation occurs under institutional control and only governed outputs are released. The talk presents the development of a shared archival infrastructure for ethical AI, called ClioX, outlining a phased approach from initial architecture and pilot implementations (2024–2025) to scaling capabilities and partnerships (2025–2026), and ultimately to decentralized, community-based governance. By shifting from disclosure to governed computation, this model enables responsible AI use while preserving sovereignty, accountability, and trust.

Victoria Lemieux is a Professor of Archival Science at the University of British Columbia’s School of Information. Additionally, she is the founder and co-lead of Blockchain@UBC, the university’s top-ranked multidisciplinary blockchain research cluster. Victoria’s research interests include risk to the privacy, security and availability of trustworthy records at the intersection of blockchain record keeping systems and artificial intelligence.

Registration for the webinar is CLOSED. Video presentation now available (members only).

 

Exploring Similarity Techniques to Support Consistent Digital Records Review

Speakers: Balint Csollei and Mark Bell, The National Archives, UK

Date: 3 June Time: 14:00 CEST.

Short introduction: Before transferring records to The National Archives, depositors must carry out sensitivity reviews to decide whether material should be opened or closed for public access. These reviews rely heavily on individual judgement, which can lead to inconsistent decisions across reviewers, organisations, and over time. When similar records are treated differently, the cumulative effect can create data mosaicking risks, where sensitive information is unintentionally revealed by combining multiple sources.

Applying exemptions consistently depends on being able to identify records with the same or similar content. Established techniques can detect duplicates and near duplicates, while newer AI approaches, such as vector embeddings, make it possible to identify conceptual similarity even where wording differs.

This presentation covers similarity-based methods including checksums, hash comparisons, near duplicate detection, and semantic similarity, and demonstrates their use through practical examples and demos to support more consistent sensitivity review.

Mark Bell is a member of the Research Team at The National Archives. He started his career at The National Archives working on the Arts and Humanities Research Council-funded Traces through Time project, researching methods for probabilistically linking biographical records. His current research interests include: the application of machine learning in the archive, particularly in the UK Government Web Archive, the use of Handwritten Text Recognition for exploring digitised collections at scale, and the implications of machine learning including Explainable AI and representing uncertainty.

Balint Csollei is an information professional with a background in digital library management. He began his career working with community media archives and now leads the government engagement team at The National Archives UK, with the aim of raising information and records management standards across departments and exploring the use of AI in appraisal and sensitivity review.

 

Registration for the webinar is CLOSED.

Back to top

 

Building evidentiary infrastructure during an active war: Lessons from the Ukraine War Archive

Speaker: Maksym Demydenko, Ukraine War Archive, Moderator: David Anderson, DLM EC Member and Treasurer

Date: 17 September Time: 11:00–12:00 CEST

Short introduction: Russia's full-scale invasion of Ukraine has become the most extensively documented war in history, generating millions of photographs, videos, witness testimonies and open-source records every day. Preserving this material is no longer simply an archival challenge: it is fundamental to future justice, historical memory and understanding contemporary wars.

In this session, Maksym Demydenko, Co-Founder and Executive Director of the Ukraine War Archive (UWA), will explore how UWA has built one of the world's largest independent digital archives documenting the war in Ukraine. Drawing on practical experience, he will discuss the challenges of preserving digital evidence during an ongoing conflict, from provenance, metadata and interoperability to AI-assisted evidence management and the ethical responsibilities of maintaining sensitive records at scale. The session will examine how archives are evolving from passive repositories into active evidentiary infrastructures that support investigators, journalists and future accountability processes. It will also consider what the Ukrainian experience can teach archivists and records professionals preparing for the digital conflicts of the future.

Maksym Demydenko is Co-Founder and Executive Director of the Ukraine War Archive (UWA), an independent non-profit organisation dedicated to preserving digital evidence of Russia's war against Ukraine. Under his leadership, UWA has developed one of the world's largest independent digital archives of the conflict, preserving hundreds of terabytes of photographs, videos, witness testimonies and open-source materials to support future accountability, historical research and public understanding. Maksym works at the intersection of digital preservation, international justice and emerging technologies, leading the development of innovative approaches to evidentiary infrastructure, AI-assisted analysis and conflict documentation. He regularly speaks on digital archives, human rights documentation and the role of technology in preserving truth during conflicts.

Registration for the webinar is CLOSED.

Back to top

 

Transforming Access to Born-Digital Archives Together

Speakers: Luís Faria, KEEP SOLUTIONS; Jan Dalsten, Danish National Archives; Mads Thilsing-Engholm, Danish National Archives

Date: 01 October Time: 13:00–14:00 CEST

Short introduction: How do you transform access to historical, born-digital data? In 2024, the Danish National Archives took a major step forward by implementing Database Preservation Toolkit (DBPTK) as their new access system—a system that has now been successfully up and running for over a year.

In this webinar, the Danish National Archives and KEEP SOLUTIONS invite you behind the scenes of their joint project. As the developers of DBPTK, KEEP led the technical implementation while developing a slew of new features tailored to modern archiving needs and together, we will talk about our joint project, showcase some of the key developments and discuss what it means for the future of DBPTK and its community.

Luis Faria is the Coordinator of the Digital Preservation Business Unit at KEEP SOLUTIONS. He is the leader of the team responsible for all RODA and DBPTK projects. He is also Director of Research and Innovation at the same portuguese company. Jan Dalsten is head of Data Science at the Danish National Archives, the department which is responsible for making born-digital records accessible for internal and external users. He is a former member of the DLM Forum Executive Committee, and has participated in numerous projects related to the development of standards and procedures regarding born-digital records. Mads Thilsing-Engholm is a special consultant at the Danish National Archives, specializing in the dissemination and accessibility of born-digital data. He has been a main actor in the development and implementation of DBPTK in the National Archives and is now administrating the system at the archive.

Registration for the webinar is OPEN.

Back to top

 

About the Sovereign Tech Agency

Speaker: Paloma Oliveira

Date: 13 October Time: 13:00–14:00 CEST

Short introduction: TBA

Registration for the webinar is OPEN.

Back to top