MITRE ATLAS adds agentic attack techniques to AI threat catalog

MITRE has updated its Adversarial Threat Landscape for Artificial-Intelligence Systems (ATLAS) knowledge base to include specific techniques targeting autonomous AI agents. These new techniques, such as AI Supply Chain Compromise (AML.T0010) and AI Agent Tool Poisoning (AML.T0110), catalog real-world attack methods used against AI systems. The update aims to provide a common vocabulary for defenders and researchers to discuss and address threats to AI agents and their supply chains. AI

IMPACT Provides a standardized vocabulary for understanding and defending against emerging threats to autonomous AI agents.

RANK_REASON Update to a structured knowledge base of adversary tactics, techniques, and procedures for AI systems. [lever_c_demoted from research: ic=1 ai=1.0]

Read on dev.to — LLM tag →

• Adversarial Threat Landscape for Artificial-Intelligence Systems
• AI Agent Tool Poisoning
• AI supply chain compromise
• AML.T0010
• AML.T0010.001
• AML.T0104
• AML.T0110
• MITRE ATLAS
• MITRE ATT&CK
• Publish Poisoned AI Agent Tool

• safety
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →