Privacy-preserving Artificial Intelligence (PEKI)

generated with ChatGPT
The German Aerospace Center (DLR) is collaborating with the University of Bonn on the HUMUS-RFWUB-PEKI project to investigate methods for privacy-preserving Artificial Intelligence (AI). The collaboration will analyse existing technologies, such as privacy-enhancing technologies, blockchain and differential privacy, to determine their practical suitability, security and performance.
In this collaborative project, the University of Bonn (RFWU) is focusing on fundamental scientific research. This includes the systematic analysis of existing AI methods with regard to their potential and limitations with respect to data protection, as well as the theoretical development of protection mechanisms such as differential privacy, federated learning and homomorphic encryption. The university is investigating the impact of these technologies on modelling accuracy, computational effort and attack resistance. Meanwhile, the RFWU is developing a reference architecture for privacy-preserving AI, implementing prototypes and conducting benchmark analyses to objectively evaluate the performance, security and scalability of different protection approaches.
Contribution Institute for AI Safety and Security
The DLR Institute for AI Safety and Security is responsible for strategic cooperation and technical validation of developed concepts. Particular focus is given to applicability in safety-critical areas.
This collaboration involves developing AI technologies that can utilise personal data without compromising user privacy. The project is researching how intelligent systems can effectively protect sensitive information. The ultimate goal is to develop robust AI applications that prioritise data protection and foster public trust in these technologies.
Institutes and facilities involved (DLR & external)
- DLR Institute for AI Safety and Security
- University of Bonn (RFWU)