OpenMined
A nonprofit developing PySyft, software for joint analysis of private data, including confidential AI evaluations.
OpenMined is a nonprofit that develops open-source software for work across private datasets 1. Its work on this map is PySyft double-blind evaluations:
- OpenMined reports that PySyft coordinated the submission, approval and execution of private code and assets in GPU enclaves for 2026 evaluations by AVERI and Singapore AISI 2.
- The pilot's technical report describes a run of Gemini 2.5 Flash Lite on private MLCommons AILuminate prompts in a Google Cloud confidential VM with an H100, using PySyft v0.10.x 3.
On this page
Implementations
Implementations this organization develops.
- PySyft coordinates an attested enclave where a model owner and evaluator run tests without sharing weights or private prompts.
Related records
Mechanisms and implementations whose records cite or describe this organization's work.
- Lets mutually distrusting parties run an agreed check over private models or records inside attested enclaves or zero-knowledge proofs, revealing only the result.
Publications
Sources this organization authored or published.
- BA. Trask et al. (2026). Double Blind Evals: Resolving the Dual Confidentiality Dilemma in AI Safety Auditing. Google DeepMind. Source recordCited by Confidential multi-party verification; PySyft double-blind evaluations; OpenMined
- COpenMined Team (2026). PySyft used for first double-blind evaluation of a proprietary, frontier-class AI model. OpenMined. Source recordCited by PySyft double-blind evaluations; OpenMined
Sources
- ANational Institute of Standards and Technology (2026). Announcement: CAISI signs CRADA with OpenMined to Enable Secure AI Evaluations. National Institute of Standards and Technology. Source recordSupports: nonprofit status, open-source software for secure work across organizations, PySyft research agreement · first paragraph
- COpenMined Team (2026). PySyft used for first double-blind evaluation of a proprietary, frontier-class AI model. OpenMined. Source recordSupports: nonprofit status, PySyft development and role in double-blind evaluations · Executive Summary; Details of the pilots; page footer
- BA. Trask et al. (2026). Double Blind Evals: Resolving the Dual Confidentiality Dilemma in AI Safety Auditing. Google DeepMind. Source recordSupports: PySyft in the 2026 pilot and OpenMined author affiliations · abstract; §2.5; §3