TRUSTEE: Trusted Research on Untrusted Systems using Trusted Execution Environments

Traditional Trusted Research Environments (TREs) often limit researchers to tightly controlled infrastructure, with limited access to advanced computing resources such as high-performance supercomputers or AI accelerators. As analytical methods become more computationally intensive, these constraints can limit the scale and ambition of research. 

While access to larger supercomputers can potentially be supported through shared responsibility models such as those being developed in the DARE UK FRIDGE Early Adopter Project, these approaches require the supercomputer administrators to be comfortable doing their bit to secure the system and data providers and TRE operators being comfortable with the supercomputer administrators sharing responsibility for securing the data.

TRUSTEE will investigate the use of Trusted Execution Environments  – an advanced technical security control provided by many modern computer processors. Trusted Execution Environments enable “Confidential Computing”, which secures data and code on expernal systems even from the administrators of those systems, enabling the secure analysis of sensitive data on external systems without needing to trust the administrators of that infrastructure. 

The project will assess technical feasibility alongside governance, assurance, and auditability requirements, focusing on how trust can be maintained beyond conventional TRE boundaries. Findings will inform future decisions about whether and how confidential computing could be integrated into national TRE infrastructure. This work will help balance flexibility, security, and public confidence.

Public involvement and engagement spans integration with DARE UK TREvolution PIE activities, a project-specific public group, and open engagement through Collaboration Cafés. This layered approach supports both deep involvement and broad dialogue around trust, security, and acceptable use.

By the end of the project, TRUSTEE will:

• Demonstrate secure analysis using Trusted Execution Environments
• Produce guidance on governance and assurance for confidential computing
• Engage diverse publics in shaping trust frameworks
• Inform future scalable TRE infrastructure design

Project information

Lead organisation: The Alan Turing Institute
Principal investigator: Dr Martin O’Reilly
Project duration: 12 months
Project partners: University College London, UK Biobank
Funding provided: £319,731
Primary contact email: moreilly@turing.ac.uk