Endpoint security

GENSEC01: How do you manage access to generative AI endpoints?

Foundation models are available for use through managed, serverless, or self-hosted endpoints. Each paradigm comes with its own security considerations and requirements. This question seeks to understand the security considerations specific to endpoints associated with generative AI workloads.

Best practices

• GENSEC01-BP01 Grant least privilege access to foundation model endpoints

• GENSEC01-BP02 Implement private network communication between foundation models and applications

• GENSEC01-BP03 Implement least privilege access permissions for foundation models accessing data stores

• GENSEC01-BP04 Implement access monitoring to generative AI services and foundation models