AWS Security Reference Architecture (AWS SRA) – AI security
AWS Security Customer Outcomes Team, Amazon Web Services (contributors)
February 2026 (document history)
| Influence the future of the AWS Security Reference Architecture (AWS SRA) by taking a short survey |
AI solutions span multiple use cases, each with distinct security requirements. The Generative AI Security Scoping Matrix
Depending on your use case, you can use a managed service where the provider handles
operations, or build your own. AWS offers a wide range of services to help you build, run,
and integrate artificial intelligence and machine learning (AI/ML) solutions of any size,
complexity, or use case. These services operate at all three layers of the generative AI stack
This guide focuses on the middle layer, which provides access to all the models and tools you need to build and scale generative AI applications and applications on AWS. Although AI (machine learning and LLMs) can be used for security purposes, this guide focuses on the foundational security controls to protect AI workloads deployed on AWS.
Intended audience
The intended audience for this guidance is security professionals, architects, and developers who are responsible for securely integrating generative AI capabilities into their organizations and applications using AWS services.
In this guide:
