Resilient Cyber w/ Christina Liaghati - Navigating Threats to AI Systems

Описание: In this episode of Resilient Cyber, host Chris Hughes sits down with Dr. Christina Liaghati, Trustworthy & Secure AI Department Manager at MITRE and the lead of MITRE ATLAS. The discussion explores navigating threats to AI systems through real-world insights and how to secure AI in the ever-evolving landscape.

Key Highlights:

-Christina’s background: Managing MITRE’s Trustworthy & Secure AI Department with a team of 55, focusing on AI assurance and security. She also leads the MITRE ATLAS project, a collaborative effort focused on AI security.

-Introduction to the MITRE ATLAS Matrix, inspired by the traditional MITRE ATT&CK Matrix. ATLAS helps characterize new AI security threats and tactics, combining AI-specific techniques with traditional cyber attack vectors.

-Discussion on AI security threats, such as data poisoning and model exploitation, and how adversaries can target AI-enabled systems as part of complex attack chains.

-Resources available from MITRE ATLAS, including the AI Security 101 page, mitigations, and threat emulation tools to help organizations navigate AI-related risks.

-Christina emphasizes the importance of standardizing AI threat characterization, enabling organizations to manage risk and security more effectively in the rapidly evolving AI landscape.

-The value of real-world case studies in grounding discussions about AI security, cutting through the hype, and providing practical insights on how organizations can address emerging threats.

-Key takeaways from tabletop exercises with CISA and industry leaders, focused on AI security incidents and the development of playbooks for information sharing and response strategies.

-Insights on the dual challenge of AI for security (using AI to bolster cybersecurity) and securing AI itself, and how the community can collaborate to address both sides of the problem.

🎧 Tune in to learn more about the MITRE ATLAS project, real-world AI security threats, and how organizations can better prepare for the future of AI security.

#AIAssurance #AISecurity #AIThreats #MITREAtlas #ChristinaLiaghati #CyberSecurity #DataPoisoning #AITrust #ThreatMitigation #AtlasProject #AISupplyChain