Watch this video on YouTube
You are responsible for implementing fine-grained access control for your AI system to ensure that only authorized users can interact with specific models and datasets. Which approach best aligns with AWS's least-privilege security principle?