All Records

Malicious Package - AI Security Technique

Malicious Package is AI attack method AML.T0011.001 with evidence level: realized. It includes 5 mitigations; 2 case studies; 1 related AI risk.

User Execution - AI Security Technique

User Execution is AI attack method AML.T0011 with evidence level: realized. It includes 5 mitigations; 2 case studies; 1 related AI risk.

AIKatz: Attacking LLM Desktop Applications - AI Case Study

AIKatz: Attacking LLM Desktop Applications - AI Case Study is an AI security case study covering type: exercise; target: LLM Desktop Applications (Claude, Ch...

Role Play Instruction

Role Play Instruction is an AI risk in 2. Privacy & Security focused on 2.2 > AI system security vulnerabilities and attacks. It is most relevant during 2 -...

Jailbreaking

Jailbreaking is an AI risk in 2. Privacy & Security focused on 2.2 > AI system security vulnerabilities and attacks. It is most relevant during 2 - Post-depl...

Discover AI Artifacts - AI Security Technique

Discover AI Artifacts is AI attack method AML.T0007 with evidence level: demonstrated. It includes 2 mitigations; 2 case studies; 5 related AI risks.

4.2 > Cyberattacks, weapon development or use, and mass harm AI Risks

4.2 > Cyberattacks, weapon development or use, and mass harm groups 72 AI risk records under 4. Malicious Actors & Misuse.

Code to Deploy Destructive AI Agent Discovered in Amazon Q VS Code Extension - AI Case Study

Code to Deploy Destructive AI Agent Discovered in Amazon Q VS Code Extension - AI Case Study is an AI security case study covering type: incident; target: Am...

Risks from models and algorithms (Risks of stealing and tampering)

Risks from models and algorithms (Risks of stealing and tampering) is an AI risk in 2. Privacy & Security focused on 2.2 > AI system security vulnerabilities...

Unsafe AI Artifacts - AI Security Technique

Unsafe AI Artifacts is AI attack method AML.T0011.000 with evidence level: realized. It includes 6 mitigations; 2 case studies; 1 related AI risk.

4.1 > Disinformation, surveillance, and influence at scale AI Risks

4.1 > Disinformation, surveillance, and influence at scale groups 74 AI risk records under 4. Malicious Actors & Misuse.

ShadowRay - AI Case Study

ShadowRay - AI Case Study is an AI security case study covering type: incident; target: Multiple systems; actor: Ray. It includes 7 case steps.

Assisting code generation for cyber security threats

Assisting code generation for cyber security threats is an AI risk in 4. Malicious Actors & Misuse focused on 4.2 > Cyberattacks, weapon development or use...

Local AI Agent - AI Security Technique

Local AI Agent is AI attack method AML.T0112.000 with evidence level: demonstrated. It includes 3 case studies; 5 related AI risks.

Modify AI Agent Configuration - AI Security Technique

Modify AI Agent Configuration is AI attack method AML.T0081 with evidence level: demonstrated. It includes 3 case studies; 5 related AI risks.

LLM Response Rendering - AI Security Technique

LLM Response Rendering is AI attack method AML.T0077 with evidence level: demonstrated. It includes 3 case studies; 5 related AI risks.

Manual Modification - AI Security Technique

Manual Modification is AI attack method AML.T0043.003 with evidence level: realized. It includes 5 mitigations; 3 case studies.

Confusing Antimalware Neural Networks - AI Case Study

Confusing Antimalware Neural Networks - AI Case Study is an AI security case study covering type: exercise; target: Kaspersky's Antimalware ML Model; actor...

7.2 > AI possessing dangerous capabilities AI Risks

7.2 > AI possessing dangerous capabilities groups 76 AI risk records under 7. AI System Safety, Failures, & Limitations.

2.1 > Compromise of privacy by leaking or correctly inferring sensitive information AI Risks

2.1 > Compromise of privacy by leaking or correctly inferring sensitive information groups 76 AI risk records under 2. Privacy & Security.