All Records

Backdoors or trojan attacks in GPAI models

Backdoors or trojan attacks in GPAI models is an AI risk in 2. Privacy & Security focused on 2.2 > AI system security vulnerabilities and attacks. It is most...

Data-related (Lack of cross-organizational documentation)

Data-related (Lack of cross-organizational documentation) is an AI risk in 7. AI System Safety, Failures, & Limitations focused on 7.3 > Lack of capability o...

Google Bard Conversation Exfiltration - AI Case Study

Google Bard Conversation Exfiltration - AI Case Study is an AI security case study covering type: exercise; target: Google Bard; actor: Embrace the Red. It i...

AI Supply Chain Reputation Inflation - AI Security Technique

AI Supply Chain Reputation Inflation is AI attack method AML.T0111 with evidence level: demonstrated. It includes 1 case study; 1 related AI risk.

Data from AI Services - AI Security Technique

Data from AI Services is AI attack method AML.T0085 with evidence level: demonstrated. It includes 5 mitigations.

Container Registry - AI Security Technique

Container Registry is AI attack method AML.T0010.004 with evidence level: demonstrated. It includes 1 case study; 1 related AI risk.

Single-User AI Agent Permissions Configuration - AI Mitigation

Single-User AI Agent Permissions Configuration is an AI mitigation for Deployment, grouped as Technical - Cyber. It is connected to 7 attack methods.

Privileged AI Agent Permissions Configuration - AI Mitigation

Privileged AI Agent Permissions Configuration is an AI mitigation for Deployment, grouped as Technical - Cyber. It is connected to 7 attack methods.

AI Bill of Materials - AI Mitigation

AI Bill of Materials is an AI mitigation for Business and Data Understanding, Data Preparation, ML Model Engineering, grouped as Policy. It is connected to 7...

Generative AI Model Alignment - AI Mitigation

Generative AI Model Alignment is an AI mitigation for ML Model Engineering, ML Model Evaluation, Deployment, grouped as Technical - ML. It is connected to 7...

Generative AI Guidelines - AI Mitigation

Generative AI Guidelines is an AI mitigation for ML Model Engineering, ML Model Evaluation, Deployment, grouped as Technical - ML. It is connected to 7 attac...

Limit Public Release of Information - AI Mitigation

Limit Public Release of Information is an AI mitigation for Business and Data Understanding, grouped as Policy. It is connected to 7 attack methods.

Facilitating fraud, scames and more targeted manipulation

Facilitating fraud, scames and more targeted manipulation is an AI risk in 4. Malicious Actors & Misuse focused on 4.3 > Fraud, scams, and targeted manipulat...

Leakage

Leakage is an AI risk in 2. Privacy & Security focused on 2.1 > Compromise of privacy by leaking or correctly inferring sensitive information. It is most rel...

Model Attacks

Model Attacks is an AI risk in 2. Privacy & Security focused on 2.2 > AI system security vulnerabilities and attacks. It is most relevant during 3 - Other.

RAG Credential Harvesting - AI Security Technique

RAG Credential Harvesting is AI attack method AML.T0082 with evidence level: demonstrated. It includes 2 mitigations; 1 case study.

Spearphishing via Social Engineering LLM - AI Security Technique

Spearphishing via Social Engineering LLM is AI attack method AML.T0052.000 with evidence level: demonstrated. It includes 2 mitigations; 1 case study.

Discover AI Model Ontology - AI Security Technique

Discover AI Model Ontology is AI attack method AML.T0013 with evidence level: demonstrated. It includes 2 mitigations; 1 case study.

Train Proxy via Gathered AI Artifacts - AI Security Technique

Train Proxy via Gathered AI Artifacts is AI attack method AML.T0005.000 with evidence level: demonstrated. It includes 2 mitigations; 1 case study.

Hacking ChatGPT’s Memories with Prompt Injection - AI Case Study

Hacking ChatGPT’s Memories with Prompt Injection - AI Case Study is an AI security case study covering type: exercise; target: OpenAI ChatGPT; actor: Embrace...