BizGuardian Connect: Privacy-Preserving Data Aggregation for Government and Business
| Reference number | |
| Coordinator | DPella AB |
| Funding from Vinnova | SEK 2 700 000 |
| Project duration | November 2023 - March 2025 |
| Status | Ongoing |
| Venture | Advanced digitalization - Enabling technologies |
| Call | Cyber security for industrial advanced digitalization 2023 |
Important results from the project
The project met its goals: a TRL 6 MVP for privacy-preserving analytics was built and validated with real data from Pulsen Omsorg; it’s now used by DPella in PoCs. A Differential Privacy API was co-developed with Chalmers and is publicly available. New techniques for balancing privacy and utility were created, resulting in a patent (PCT/SE2024/050923) and top-tier publication. The MVP was tested with two municipalities, leading to an on-premise pivot and a privacy risk assessment.
Expected long term effects
The project has long-term impact on privacy tech and policy. It advanced Differential Privacy (DP) by improving methods to balance privacy and accuracy, key for complex analytics. A standardized API and on-premise solution ease adoption. AI-based de-anonymization experiments showed how tools like ChatGPT can infer sensitive info, challenging traditional anonymization and highlighting the urgent need for strong, science-backed methods like DP.
Approach and implementation
The project followed the plan with the right activities and expected results. It also led to valuable insights: testing revealed scalability limits in DP tools, prompting work on relational database integration; concrete use cases showed that aggregated data can leak sensitive data and be re-identified by AI; and new techniques to estimate accuracy in ratios resulted in a patent. A delay due to team changes occurred, but collaboration exceeded expectations, enabling discoveries and follow-ups.