Artificial intelligence for oral history (KI.OH). Development and evaluation of AI tools for the curation and reuse of sensitive audiovisual research data
The project “Artificial Intelligence for Oral History” (KI.OH) develops and evaluates a portfolio of AI-supported applications for the annotation and reuse of sensitive audiovisual research data. Using the use case of oral history, it examines the feasibility and connectivity, the potential and risks of AI-based curation, research, and trust tools that will also be useful for other infrastructures.
Curation tools are being developed to assist in the annotation and sharing of interviews, particularly in transcription using speech recognition, in entity recognition and keyword assignment in accordance with authority files, and in summarization. In addition, there are consulting and search systems that support reuse with methodological advice from a digital research assistant and with semantic search across transcripts and other interview content.
The risks associated with the use of AI are reflected upon with AI experts and also addressed technically with a prototypical trust level engine. In exchange with data providers and data users, requirements are collected, workflows specified, prototypes evaluated, and results publicized. The tool portfolio is being developed in an interoperable system architecture so that it is sustainable and connectable for other research communities beyond oral history.
The project team combines many years of expertise in oral history, particularly in the development of interview archives and indexing platforms (Freie Universität Berlin and FernUniversität Hagen), with in-depth knowledge of AI development (Technische Informationsbibliothek Hannover) and trust research (University of Cologne), as well as internationally established research on digital humanities and critical AI (Technical University of Darmstadt).
Duration and Funding- 1.1.2027 - 31.12.2029
- Deutsche Forschungsgemeinschaft (DFG), Scientific Library Services and Information Systems Research data and software
- https://gepris.dfg.de/gepris/projekt/577909525?lang=en
- Professor Dr. Øyvind Eide, Susanne Kurz, Universität zu Köln
- Dr. Almut Leh, Dr. Dennis Möbus, FernUniversität in Hagen
- Professorin Julianne Nyhan, Ph.D., Technische Universität Darmstadt
- Dr. Jennifer D' Souza, Technische Informationsbibliothek Hannover
Freie Universität Berlin
University Library
Digital Interview Colelctions
Project KI.OH
Garystr. 39
14195 Berlin
Web: https://www.fu-berlin.de/sites/ub/forschen/interviewsammlungen/forschung/kioh
Mail: ki.oh@oral-history.digital