We’re proud to share that OpenMined’s work has been highlighted in the newly released United Nations Guide on Privacy-Enhancing Technologies for Official Statistics (UN PET Guide). The Guide showcases leading examples of privacy-preserving technologies being deployed across government, industry, and academia.
One of the featured case studies details our ongoing partnership with Twitter to evaluate PETs for algorithmic transparency. The project aims to enable external researchers to study proprietary data and models held by Twitter without having direct access to the data. Using OpenMined’s PySyft software for remote execution and differential privacy, external researchers can reproduce and validate Twitter’s internal research findings whilst protecting the privacy of the data. This groundbreaking work has led to the Christchurch Call Initiative on Algorithmic Outcomes – a partnership between New Zealand, the United States, Twitter, Microsoft, and OpenMined to develop privacy-preserving research capabilities across multiple platforms.
The guide also highlights OpenMined’s collaboration with the UN PET Lab on international trade analysis. Working with multiple national statistical offices, including Statistics Canada, the US Census Bureau, the UK Office for National Statistics, Statistics Netherlands, and ISTAT Italy, we developed privacy-preserving methods for analyzing trade data across countries. This pilot demonstrates how PETs can enable countries to identify asymmetries in trade information – something they were previously unable to do due to constraints on sharing such information with one another. The project demonstrates how PySyft’s federated data network can enable joint statistical analysis while protecting sensitive national trade information.
OpenMined’s Executive Director, Andrew Trask, had the distinct pleasure to serve as a co-author of Chapter 2: Methodologies and Approaches. This chapter provides detailed technical guidance on key PETs, including federated learning, differential privacy, and secure multi-party computation.
We’re honored to have our work recognized in this important UN publication and are excited to contribute to advancing the adoption of privacy-preserving technologies in official statistics. The Guide represents a significant milestone in establishing best practices for protecting privacy while enabling valuable data analysis.
You can read the UN PET Guide here: https://unstats.un.org/bigdata/task-teams/privacy/guide/2023_UN%20PET%20Guide.pdf
