AI Security Maturity Model·AWS v1.0.0

Drag to explore · pinch to zoom
Home / Matrix / Capability 6 / c6.4
Enhanced AI Application SecurityDefense-in-depth layer: AI Application Capability 6 · AI applicationsUse case: AI that answers 👤 CustomerAWS Shared Responsibility Model: who implements and operates this control. Customer = you own it; AWS = AWS operates it and you inherit it; Shared = both have obligations.

c6.4 Protect the app interface with WAF and TLS

Front web-based AI application interfaces with AWS WAF and enforce TLS, so common web exploits, bot traffic, and prompt-injection patterns are blocked at the edge.

Foundational
Enhanced
Advanced
Effort: Medium Impact: High GenAI scopesFrom the AWS Generative AI Security Scoping Matrix.: 3 · Pre-trained models

Implementation guidance

How to implement on AWS

Deploy AWS WAF in front of the AI app interface with managed rules, custom prompt-injection detection rules, and rate limiting. Enforce TLS 1.2+ for all user connections and use AWS Certificate Manager for certificate issuance and rotation.

Acceptance criteria

WAF protects the interface with managed + custom rules and rate limiting; all user connections use TLS 1.2+ with ACM-managed certificates.

Practical implementation

How this control is implemented in each reference architecture:

AWS WAF fronts the AI app interface in the Application account with managed + custom rules and rate limiting, and ACM issues/rotates the TLS 1.2+ certificates.

  1. Deploy AWS WAF in front of the AI app interface with managed rules and custom prompt-injection detection rules.
    Application account AWS WAF
  2. Add rate-based rules to stop bot traffic and prompt stuffing.
    Application account AWS WAF
  3. Enforce TLS 1.2+ for all user connections using AWS Certificate Manager for issuance and rotation.
    Application account Certificate Manager

How to validate: WAF protects the interface with managed + custom rules and rate limiting; all user connections use TLS 1.2+ with ACM-managed certificates.

Avoid: A public AI chat UI with no WAF or rate limits. Bots and injection hit it immediately; protect the edge and enforce TLS.

Generative AI OU (multi-account) →

WAF with managed + custom rules and rate limiting protects the app interface in the account, with ACM-managed TLS 1.2+.

  1. Deploy AWS WAF in front of the interface with managed rules and custom prompt-injection rules.
    Application tier (IAM role) AWS WAF
  2. Add rate limiting for bot/prompt-stuffing traffic.
    Application tier (IAM role) AWS WAF
  3. Enforce TLS 1.2+ with ACM-managed certificates.
    Application tier (IAM role) Certificate Manager

How to validate: WAF protects the interface with rate limiting; all connections use TLS 1.2+ with ACM certificates.

Avoid: Serving the AI app over plain HTTP or without WAF. Enforce TLS and filter at the edge.

Single account (entry point) →

Threats addressed

  • Web exploits (injection, XSS), bot traffic, and prompt stuffing against the app.
  • Resource exhaustion from unthrottled requests.

AWS services

AWS WAF AWS Certificate Manager

Framework mappings

OWASP GenAI/LLM Top 10 (2026)

  • LLM01 Prompt Injection
  • LLM06 Unbounded Consumption

NIST AI Risk Management Framework (AI 100-1)

  • MANAGE.1 AI risks are prioritized, responded to and managed

AWS Well-Architected Generative AI Lens (Security)

  • GENSEC04-BP02 Sanitize and validate user inputs to foundation models

Learning resources

References

  • AWS SRA for AI — Capability 6, Network security