ESRA 2027 sessions by theme
Back to Overview of Sessions
Using AI for Working With Sensitive Data – Needs and Potential Solutions |
| Coordinator 1 | Dr Frederic Gerdon (GESIS – Leibniz Institute for the Social Sciences) |
| Coordinator 2 | Dr István László Gyimes (Social Sciences and Digital Humanities Archive (SODHA) State Archives of Belgium / Vrije Universiteit Brussel) |
| Coordinator 3 | Beate Lichtwardt (UK Data Service / UK Data Archive, University of Essex) |
| Coordinator 4 | Dr Deborah Wiltshire (GESIS – Leibniz Institute for the Social Sciences) |
| Coordinator 5 | Dr Lorraine Wong (CESSDA ERIC) |
Session Details
“AI”-based tools hold promise for collecting, accessing, or analyzing sensitive data (such as georeferenced surveys or many types of qualitative data) more efficiently, thereby increasing capacities for complex analyses and making the lives of researchers and data managers easier. However, processing data this way requires heightened safeguards to protect confidentiality. Researchers and data managers thus need to grapple with two challenges at the same time: they need to acquire the necessary skills to make sensible use of AI-based tools (1) for handling research data and (2) to work with sensitive data in a safe way. These two challenges combine into the hitherto barely explored terrain of defining and applying the safeguards needed to responsibly use AI-based tools for sensitive data. There are several pathways to implement safe data access and analysis, and they need to be tailored to the researchers’ needs. Safeguards may include training, governance frameworks, and technical security setups. Trusted Research Environments (TREs), for instance, are a technical security setup that can put such safeguards into practice: they offer restricted and safe (remote or on-site) conditions for providing data access for researchers.
In this session, we bring together perspectives from researchers (including less and more computationally orientated social scientists) and data management professionals to discuss the needs of different user groups when it comes to sensible and productive use of AI-based tools for working with sensitive data. Topics of interest for contributions to this session include, but are not limited to:
- Experiences/case studies of implementing AI-based tools when working with sensitive data
- Needs and challenges of researchers using AI-based tools with sensitive data, including navigating restrictions on data sharing and access
- Governance, guidelines, and infrastructures for enabling responsible AI use with sensitive data, including standards, access solutions, and risk assessments
