Privacy Enhancing Technologies for Official Statistics (PET4OS)

This page gathers resources and information about innovation-oriented activities carried out by Eurostat in the field of Privacy Enhancing Technologies for Official Statistics (PET4OS).

Announcements

• On 13/3/2024 the European Parliament approved an amendment to the Regulation (EC) 223/2009 on European Statistics. The adopted text is available here. The agreed legal text makes explicit reference to Privacy Enhancing Technologies and to "secure infrastructure [...] based on technologies that are specifically designed to comply with [GDPR and EUDPR]".  This follows the adoption by the European Commission of a proposal for revising the European statistics Regulation in order to make it "fit for the future" and "tap the potential of new data sources". The EDPS opinion on the EC proposal was published on 6/9/2023 and is available here. 
• On 04/2023 Eurostat published an open call for tender with subject:  Specification, feasibility analysis and prototype demonstration of a multi-party secure private computing system for processing confidential sets of micro-data across organisations in support of statistical innovation (TSS-PET). The submission deadline closed on 31/7/2023. The call documents are available here. 
• On 20/1/2023 the European Commission adopted a proposal for a new EU regulation on European statistics on population and housing (ESOP) making explicit reference to Privacy Enhancing Technologies in the (proposed) legal text - see specifically Article 13, Article 14 and Recital (30). The EDPS opinion was published on the ESOP proposal was published on 16/3/2023 and is available here.  

Running Projects

• JOCONDE - Joint On-demand COmputation with No Data Exchange. 

Past Projects

• Project ESTAT.2019.0232. A project exploring the use of Trusted Execution Environment (TEE) with hardware isolation for the processing of Mobile Network Operator (MNO) data for official statistics. The project closed in 2021. All deliverables and project reports are publicly available from the project page. 
• UNECE HLG-MOS Input Privacy-Preservation project. The Final Report for the project is available here. As part of this project, an open technical consultation was launched on the concept of MultiParty Secure Private Computing-as-a-Service (MPSPCaaS). 

Papers

• A Cautionary Reflection on (Pseudo-)Synthetic Data from Deep Learning on Personal Data, short paper presented at Privacy in Statistical Databases  (PSD2024) conference, Antibes, 25-27 September 2024 (slides). 
• A reflection on privacy and data confidentiality in Official Statistics. Paper presented at the 62nd ISI World Statistics Conference, Kuala Lumpur, August 2019.
• Privacy and data confidentiality for Official Statistics: new challenges and new tools. Paper at NTTS'19 conference, Brussels, 11-15 March 2019.
• Trusted Smart Surveys: a possible application of Privacy Enhancing Technologies in Official Statistics, 2020. Short paper accepted at SIS2020 meeting, Pisa, 22-24 June 2020. 
• Trusted Smart Statistics: How new data will change official statistics. Paper published in Data & Policy journal, 2020, discussing  the prospective role of Secure Private Computing in fostering new approaches to official statistics production. 
• Trusted smart statistics: Motivations and principles. Paper published in the Scientific Journal of the IAOS, 2019. 

Presentations

• When GDPR met COED … (a tale of two made one for the other) - Talk given at the  Computing over Encrypted Data (COED) Industry Day, KU Leuven,  10/10/2023. 
• Towards a shared European Statistical System infrastructure for collaborative confidential computing - Talk given an INEXDA meeting 1/12/2023.
• Beyond the lab: How Multi-Party Secure Private Computing-as-a-service (MPSPCaaS) can foster the adoption of Input Privacy technologies in the ESS - Talk given at the  Working Group on Methodology meeting,  28/3/2023. 
• A reflection on the potential role of MultiParty Computation for the production of (future) Official Statistics - talk at ENISA Workshop on Personal Data Sharing - Emerging Technologies 7/10/2022. 
• A reflection on the potential role of MultiParty Computation for the production of (future) Official Statistics - keynote talk at MPC Data Privacy Summit 5/10/2022. 
• Input Privacy vs. Output Privacy: a reflection from the perspective of official statistics - Workshop on Data Protection Engineering: Integrating safeguards into the processing, Garante per la Protezione dei Dati 21/9/2022. 
• Towards a reference architecture for Trusted Smart Surveys - Meeting of UNECE HLG-MOS Input Privacy-Preservation Project 22/1/2021.
• Access to external data for the production of official statistics and personal data protection: what can we demand from Secure Private Computing technologies? Talk at ISTAT workshop on personal data protection 23/1/2021.
• The role of Privacy Enhancing Technologies in future Trusted Smart Statistics. Presentation given at Cyber-Defence Campus Conference 2020 (CYD'20).