QuestionQ56

VerifiedDeploy AI-powered business solutions

A company has an AI solution called Solution1 deployed in the production environment. Solution1 uses an Azure OpenAI model to generate marketing emails for existing customers.

During an internal review, you discover that Solution1 generates different emails based on customers’ traits.

You need to recommend a bias-mitigation strategy. The strategy must comply with Microsoft responsible AI principles.

What should you recommend?

Explanation

System instructions can steer an Azure OpenAI model’s behavior and add explicit safety and fairness constraints. Prompt engineering, including metaprompt or system-message tuning, is an application-level mitigation Microsoft recommends for potential Responsible AI harms; instructions can require consistent treatment of customers and prohibit differential content based on protected or otherwise inappropriate traits.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!