Summit: TrustWorthy Ai
Digital Futures: Health, Technology, and Cities
march 20, 2023
Trustworthy AI is a programme of ITU (United Nations Specialized Agency for ICT), which advances the standardization of several key Privacy-enhancing technologies including homomorphic encryption, federated learning, multi-party computation, differential privacy and zero-knowledge proofs.
Privacy-Enhancing Technologies are complex and their workings can be considered counterintuitive. Homomorphic encryption enciphers data, while allowing computation. Other PETs exhibit similar characteristics. For the adoption of PETs, it is essential to prove to potential users that no data is revealed. Often these technologies are developed with the explicit goal of enabling their analysis across jurisdictions (compliant e.g. GDPR). Standard creation needs to involve government, academia, and the private sector. The programme developed out of the ITU-WHO Focus Group on Artificial Intelligence for Health FG-AI4H), which revealed the need for international collaboration on data analysis, while respecting privacy.