Public Report: Building Trust in AI Through a Cyber Risk-Based Approach

This February 2025 document, jointly published by the French National Cybersecurity Authority (ANSSI) and numerous international partners, presents a high-level analysis of cybersecurity risks associated with Artificial Intelligence (AI) systems. It emphasizes a risk-based approach to building trust in AI, focusing on vulnerabilities throughout the AI lifecycle, from data collection to deployment. The report highlights key attack vectors, including data poisoning, model extraction, and evasion techniques, and offers guidelines for developers, users, and policymakers to mitigate these risks. The document also stresses the importance of considering AI supply chain security and the need for ongoing international collaboration to address the evolving threat landscape. Finally, it includes checklists and recommendations for secure AI system implementation.

https://cyber.gouv.fr/en/publications/building-trust-ai-through-cyber-risk-based-approach

https://www.cyber.gc.ca/en/news-events/joint-guidance-building-trust-artificial-intelligence-through-cyber-risk-based-approach