MITRE ATLAS (Adversarial Threat Landscape for Artificial-Intelligence Systems) is a comprehensive, publicly available knowledge base and framework that documents the tactics, techniques, and case studies of adversarial attacks targeting artificial intelligence (AI) and machine learning (ML) systems. Modeled after the MITRE ATT&CK framework, ATLAS organizes real-world observations of how adversaries compromise AI systems, including methods like data poisoning, model evasion, and model inversion. The framework categorizes 14 main adversarial tactics and provides detailed examples of techniques used to exploit AI vulnerabilities, helping organizations anticipate, detect, and defend against these evolving threats. ATLAS serves as a critical resource for security professionals, AI developers, and researchers to strengthen AI security and foster collaboration in the face of rapidly advancing AI-driven risks.
Publication's URL
https://atlas.mitre.org/Additional documents on this topic
- OWASP Practical Guide for Secure MCP Server Development
- NIST IR 8596 Cybersecurity Framework Profile for Artificial Intelligence (Cyber AI Profile): NIST Community Profile ★★★★★
- ETSI EN 304 223: Securing Artificial Intelligence (SAI) – Baseline Cyber Security Requirements for AI Models and Systems
- CSA Data Security within AI Environments
- CISA Principles for the Secure Integration of Artificial Intelligence in Operational Technology