AI-900 Practice Question: Describe features of generative AI workloads on Azure
A company uses a GPT-based model to generate marketing copy. They notice the model occasionally produces text that includes harmful stereotypes. They want to reduce these harmful outputs without retraining the model. Which approach is most appropriate?
⚠ Common exam trap
It's easy for candidates to confuse fine-tuning (which requires retraining) with prompt engineering (which does not), or assume that adjusting parameters like temperature or max tokens can fix content quality issues, when in fact they only affect randomness and length, not semantic safety.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Prompt engineering with specific instructions to avoid stereotypes
Prompt engineering allows you to guide the model's behavior at inference time without modifying its weights. By including explicit instructions in the prompt (e.g., 'Avoid harmful stereotypes'), you can steer the output toward safer content. This is the most appropriate approach when retraining is not an option, as it directly addresses the undesired outputs through input design.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Fine-tuning the model on a curated dataset
Why it's wrong here
Fine-tuning would update the model's weights by running extra gradient optimization steps on a curated marketing dataset, which is a form of retraining. Since the user explicitly wants to avoid retraining the model, this approach is not allowed. It also requires a labeled dataset and substantial compute, and the curated data may inadvertently reinforce subset-specific bias rather than removing global stereotypes.
- ✓
Prompt engineering with specific instructions to avoid stereotypes
Why this is correct
Prompt engineering modifies only the input context at inference time, adding explicit constraints such as 'Do not generate stereotypes about gender, race, or age' to steer the model's next-token probabilities. Because GPT models are trained to follow instructions, these directives can effectively suppress biased outputs without touching the frozen model weights. This makes it a fast, lightweight, and iterable safety control.
- ✗
Reducing the temperature parameter to zero
Why it's wrong here
Temperature scaling adjusts the softmax distribution over the vocabulary; setting it to zero forces greedy token selection, making outputs deterministic. However, deterministic sampling does not change the underlying learned associations in the transformer weights, so a stereotype that is the most probable completion will still be produced. It also harms output quality by eliminating all variation.
- ✗
Increasing the maximum output length
Why it's wrong here
Increasing the maximum output length raises the token-generation stop threshold, giving the model more decoding steps to continue unconstrained generation. This does not introduce any content-level filtering, so the model may wander into a longer sequence that repeats or amplifies stereotyped tropes from training data. The effect is additive: more tokens mean more opportunities for unsafe content to surface.
Go deeper
Related to this question
Learn chapter
Prompt Engineering Fundamentals
Key term
Prompt engineering
Prompt engineering is the practice of designing and refining input queries to AI models to get the most accurate, relevant, and useful outputs.
Key term
Model
In IT and AI, a model is a trained mathematical representation that learns patterns from data to make predictions or decisions.
About these practice questions
One of 985 original AI-900 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.