cPAID

The cPAID project, funded by Horizon Europe, is at the forefront of enhancing the security and resilience of AI applications and operations. It is designed to research, develop, and deliver an advanced, cloud-based, platform-agnostic defense framework that provides holistic protection for AI systems against adversarial attacks and malicious actions.

As AI becomes increasingly embedded in the fabric of organizations, securing these systems from evolving threats is more critical than ever. cPAID tackles this challenge by focusing on both poisoning and evasion adversarial attacks, combining state-of-the-art AI defense methods such as life-long semi-supervised reinforcement learning, transfer learning, adversarial training, and feature reduction. These methods will be bolstered by integrating security- and privacy-by-design principles, privacy-preserving techniques, explainable AI (XAI), Generative AI, context-awareness, as well as risk and vulnerability assessments. Additionally, the project will leverage threat intelligence to enhance the protection of AI systems.