Open.Oral-History. Recommendations and tools for risk assessment, anonymization, and the provision of legally protected and ethically sensitive audiovisual interviews.

• Applicants: Dr. Almut Leh; Dr. Dennis Mischke
• Subject Area: Modern and Contemporary History
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 554898678

Biographical narrative interviews of oral history are subject to specific requirements of ethical and legal protection. They are more difficult to anonymize than other research data because they contain acoustic and visual information. Anonymization must prevent the re-identification of individuals. At the same time, the source must be altered as little as possible in the interest of qualitative research. Therefore, there are considerable legal, ethical, and subject-specific barriers to making them available in accordance with according to the FAIR Principles. At the same time, there is significant interdisciplinary research interest in their subsequent use. Thus, they are a particularly challenging example of (still) copyright-protected objects. Collection managers lack a structured knowledge of the legal and technical requirements for digital distribution. The complexities of data protection, privacy, and copyright law, such as missing or inadequate consent forms or the need for anonymization, are often a seemingly insurmountable hurdle. The “Open.Oral-History” project addresses this need by helping collection managers to systematically assess the requirements for making their interviews available and, if necessary, to enable anonymization using an artificial intelligence (AI)-assisted anonymization process. Together with the managers of interview collections in German, English, Polish and Ukrainian, exemplary use cases will be identified. With legal support, workflows for legal clarification, risk assessment and publication will be developed for these different scenarios. The results will be published as a handbook. For cases in which anonymization allows the access to the interviews, generic solutions for the legally compliant provision of the interviews will be developed. AI-based speech recognition will be used to automatically generate machine-readable transcripts from which words and passages to be anonymized will be identified and automatically masked in the audio and/or video. Collection managers receive the anonymized text derivatives and audiovisual media, as well as machine-readable license and rights information for the provision and re-use of the oral history interviews. The legally reviewed handbook and the prototypical technological procedures that are developed should be generically applicable to data sets from other disciplines. The transferability of the developed solutions will be determined in exchange with other pilot projects and representatives of related disciplines.

• DFG Programme: Cataloguing and Digitisation (Scientific Library Services and Information Systems)
• Co-Investigators: Dr. Tobias Kilgus; Verena Lucia Nägel; Dr. Cord Pagenstecher