Supporting Safe Synthetic Data Release: UK Data Service Workflows, Standards, and Community Practice

Synthetic data is increasingly used to support research. It can help researchers learn data structures, develop code, and explore feasibility for their research projects. But synthetic data only becomes truly useful when it is FAIR, including being curated, documented, quality-checked, and disseminated in ways researchers can trust and reuse.

This presentation shares how the UK Data Service supports the provision of synthetic data as part of responsible access and researcher enablement. Building on earlier work, including the UKDS Balancing the Data Scales project on the costs and benefits of low-fidelity synthetic data, we focus on the often-invisible “plumbing” that makes synthetic data usable in practice: producer support, packaging and release workflows, supporting technical infrastructure, training and guidance, including a minimal documentation standard that promotes transparency and reproducibility for synthetic data collections.

We also situate this work within sector collaboration, to the UK-wide synthetic data community emerging through DARE UK Synthetic Data Community Groups and the ADR UK Synthetic Data Working Group, and other international collaborations such as FAIR AI Data.