PromptRiskDBThreat intelligence atlas
AI Risk

Security

"Encompasses vulnerabilities in AI systems that compromise their integrity, availability, or confidentiality. Security breaches could result in significant harm, ranging from flawed decision-making to data leaks. Of special concern is leakage of AI model weights, which could exacerbate other risk areas."

View related techniquesRead profile
AI Risk2. Privacy & Security2.2 > AI system security vulnerabilities and attacks2 - Post-deployment

Record summary

A quick snapshot of what this page covers.

Techniques1Attack methods connected to this risk.
Mitigations0Defenses that may help with related attacks.
Domain2. Privacy & SecurityThe broad risk area this belongs to.

Risk profile

How this risk is described and categorized.

Domain2. Privacy & Security
Subdomain2.2 > AI system security vulnerabilities and attacks
Entity2 - AI
Intent2 - Unintentional
Timing2 - Post-deployment
CategorySecurity
Subcategoryn/a

Related techniques

Attack methods connected to this risk.

Suggested mitigations

Defenses that may help with related attacks.

No propagated mitigations. No defense is available through the connected attack methods.

Source

Research source for this risk, when available.