All Records

LLMSmith: RCE Vulnerabilities in LLM-Integrated Applications - AI Case Study

LLMSmith: RCE Vulnerabilities in LLM-Integrated Applications - AI Case Study is an AI security case study covering type: exercise; target: LLM Integration Fr...

Achieving Code Execution in MathGPT via Prompt Injection - AI Case Study

Achieving Code Execution in MathGPT via Prompt Injection - AI Case Study is an AI security case study covering type: exercise; target: MathGPT (https://mathg...

Attack on Machine Translation Services - AI Case Study

Attack on Machine Translation Services - AI Case Study is an AI security case study covering type: exercise; target: Google Translate, Bing Translator, Systr...

Data from Local System - AI Security Technique

Data from Local System is AI attack method AML.T0037 with evidence level: realized. It includes 3 case studies; 5 related AI risks.

Model - AI Security Technique

Model is AI attack method AML.T0010.003 with evidence level: realized. It includes 5 mitigations; 4 case studies.

AI Telemetry Logging - AI Mitigation

AI Telemetry Logging is an AI mitigation for Deployment, Monitoring and Maintenance, grouped as Technical - Cyber. It is connected to 17 attack methods.

Cybersecurity

Cybersecurity is an AI risk in 4. Malicious Actors & Misuse focused on 4.3 > Fraud, scams, and targeted manipulation. It is most relevant during 2 - Post-dep...

Full AI Model Access - AI Security Technique

Full AI Model Access is AI attack method AML.T0044 with evidence level: demonstrated. It includes 2 mitigations; 3 case studies; 6 related AI risks.

Prompt Leaking

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

Backdoor Attack on Deep Learning Models in Mobile Apps - AI Case Study

Backdoor Attack on Deep Learning Models in Mobile Apps - AI Case Study is an AI security case study covering type: exercise; target: ML-based Android Apps; a...

Privacy and consent

Privacy and consent is an AI risk in 4. Malicious Actors & Misuse focused on 4.3 > Fraud, scams, and targeted manipulation. It is most relevant during 2 - Po...

Publish Poisoned Datasets - AI Security Technique

Publish Poisoned Datasets is AI attack method AML.T0019 with evidence level: demonstrated. It includes 3 mitigations; 1 case study; 11 related AI risks.

Financial Transaction Hijacking with M365 Copilot as an Insider - AI Case Study

Financial Transaction Hijacking with M365 Copilot as an Insider - AI Case Study is an AI security case study covering type: exercise; target: Microsoft 365 C...

Delay Execution of LLM Instructions - AI Security Technique

Delay Execution of LLM Instructions is AI attack method AML.T0094 with evidence level: demonstrated. It includes 1 case study; 12 related AI risks.

Call Chains - AI Security Technique

Call Chains is AI attack method AML.T0084.003 with evidence level: demonstrated. It includes 1 case study; 12 related AI risks.

Gather RAG-Indexed Targets - AI Security Technique

Gather RAG-Indexed Targets is AI attack method AML.T0064 with evidence level: demonstrated. It includes 1 case study; 13 related AI risks.

AI Software - AI Security Technique

AI Software is AI attack method AML.T0010.001 with evidence level: realized. It includes 2 mitigations; 5 case studies; 1 related AI risk.

Erode AI Model Integrity - AI Security Technique

Erode AI Model Integrity is AI attack method AML.T0031 with evidence level: realized. It includes 4 mitigations; 5 case studies.

Data Exfiltration via Agent Tools in Copilot Studio - AI Case Study

Data Exfiltration via Agent Tools in Copilot Studio - AI Case Study is an AI security case study covering type: exercise; target: Copilot Studio Customer Ser...

Transferable adversarial attacks from open to closed-source mod- els

Transferable adversarial attacks from open to closed-source mod- els is an AI risk in 2. Privacy & Security focused on 2.2 > AI system security vulnerabiliti...