AB-100 Certification Practice Question #112

Question

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

During an internal review, you identify that Solution1 creates different emails depending on the customers’ traits.

You need to recommend a strategy to mitigate the bias. The strategy must adhere to Microsoft responsible AI principles.

What should you recommend?

Accepted Answer

The scenario involves a deployed production Azure OpenAI-based solution creating trait-differentiated emails. The constraint is Microsoft responsible AI principles. Two viable answers compete: B (modify system instructions) is the immediate, practical mitigation — the system prompt can explicitly instruct the model to generate trait-neutral marketing emails regardless of customer attributes, and MS Learn confirms "metaprompt tuning can be an effective mitigation for many types of harm." D (modify training dataset) addresses root-cause bias but is infeasible for a deployed Azure OpenAI commercial model where the organization cannot alter the underlying training data. Since Solution1 uses Azure OpenAI models (not a custom-trained model), modifying the training dataset is not actionable. The suggested answer B and 24b8eec's reasoning align with the practical, exam-context-appropriate answer. Confidence is moderate due to the genuine D vs. B split.

More AB-100 practice questions