AI Threat Model

Assessment SECURE

Purpose

Structured threat modeling template for AI systems covering attack surfaces, threat actors, and mitigation strategies.

Related Controls

ISO A.8 NIST AI-600-1

1. System Overview

Describe the AI system being threat modeled, its components, and trust boundaries.

System Name: [SYSTEM NAME]

Version: [VERSION]

Date: [DATE]

Threat Modeler: [NAME], [ROLE TITLE]

System Description: [Brief description of what the system does, its architecture, and key components]

Components:

  • Frontend/UI: [Description]
  • API Layer: [Description]
  • AI Model/Service: [Description]
  • Data Store: [Description]
  • External Integrations: [Description]

Trust Boundaries:

  • User ↔ Application
  • Application ↔ AI Service
  • AI Service ↔ Data Store
  • Application ↔ External APIs

2. Threat Actor Profiles

Identify who might attack this system and their capabilities.

ActorMotivationCapabilityAccess Level
External AttackerData theft, disruption, financial gainPrompt injection, API abuse, social engineeringUnauthenticated / public endpoints
Malicious UserBypass restrictions, extract data, abuse servicePrompt manipulation, jailbreaking, rate limit abuseAuthenticated user
Compromised InsiderData exfiltration, sabotageDirect system access, model poisoning, configuration changesPrivileged access
Supply ChainBackdoor, data collectionCompromised model, poisoned training data, malicious dependenciesVendor/integration level
Automated AgentResource exhaustion, data harvestingScripted attacks, bot networks, recursive promptsAPI access

3. Attack Surface Analysis

Enumerate AI-specific attack vectors using STRIDE or similar methodology.

Attack VectorSTRIDE CategoryDescriptionLikelihoodImpactRisk
Prompt Injection (Direct)TamperingUser crafts input to override system instructions
Prompt Injection (Indirect)TamperingMalicious content in retrieved data manipulates model behavior
Training Data PoisoningTamperingAdversary corrupts training/fine-tuning data
Model InversionInfo DisclosureAttacker extracts training data from model responses
Sensitive Data LeakageInfo DisclosureModel reveals PII, credentials, or proprietary data
Denial of ServiceDenial of ServiceResource exhaustion via large/recursive prompts
JailbreakingElevationBypass content filters and safety guardrails
Agent HijackingElevationManipulate AI agent to perform unauthorized actions

4. Mitigations

For each identified threat, document the mitigation strategy and its current status.

ThreatMitigationControl TypeStatusOwner
Prompt InjectionInput sanitization, system prompt hardening, output filteringPreventive
Data LeakageOutput scanning for PII/secrets, data classification enforcementDetective
DoSRate limiting, token limits, request size limits, timeout enforcementPreventive
Model TheftAccess controls, API key rotation, usage monitoringPreventive
JailbreakingContent filters, output validation, monitoring for policy violationsDetective
Agent HijackingLeast privilege, tool whitelisting, human approval gatesPreventive

Control Types: Preventive (stop the attack), Detective (detect the attack), Corrective (respond to the attack)

5. Residual Risk & Sign-Off

Document remaining risks after mitigations and get formal approval.

Residual Risks

  1. [RISK — e.g., "Novel prompt injection techniques may bypass current filters"]

- Likelihood after mitigation: [Low/Medium/High]

- Accepted by: [NAME, TITLE]

  1. [RISK — e.g., "Model may generate plausible but incorrect outputs"]

- Likelihood after mitigation: [Low/Medium/High]

- Accepted by: [NAME, TITLE]

Approval

  • Threat Model Review: [NAME] — [DATE] — Approved / Requires Revision
  • Security Sign-Off: [NAME] — [DATE]
  • System Owner Acceptance: [NAME] — [DATE]

Next Review

  • Scheduled: [DATE] (or triggered by significant system change, new threat intelligence, or security incident)
← Back to all templates