AI-900 Practice Question: Describe features of generative AI workloads on Azure
What is 'guardrails' in generative AI applications and how are they implemented?
⚠ Common exam trap
Watch out — candidates often confuse operational controls (rate limits) or legal agreements (terms of service) with technical safety mechanisms (guardrails), which are specifically designed to filter and validate AI outputs in real time.
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
✓
Safety and quality constraints (content filters, system prompts, output validation) preventing harmful AI outputs
Guardrails in generative AI applications are safety and quality constraints implemented to prevent harmful or inappropriate AI outputs. They include content filters that block offensive language, system prompts that steer model behavior, and output validation that checks responses against predefined policies. This is correct because guardrails are a core feature of responsible AI deployment, ensuring that generative models like GPT-4 in Azure OpenAI Service produce safe, compliant, and contextually appropriate content.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Physical barriers around AI data centres to prevent unauthorised access
Why it's wrong here
Physical barriers secure data centre facilities and hardware, but they are infrastructure-level security controls, not guardrails. Guardrails are software-based safety mechanisms that operate at runtime within the AI system, filtering prompts and outputs, and constraining model behaviour through prompts and validation. A physical barrier cannot prevent a model from generating harmful text — it only prevents physical intrusion.
- ✓
Safety and quality constraints (content filters, system prompts, output validation) preventing harmful AI outputs
Why this is correct
Guardrails in Azure AI refer to a layered set of safety and quality controls: content filters (e.g., Microsoft's safety classifiers for hate, violence, sexual, self-harm), system prompts that steer model tone and scope, and output validation (e.g., grounding checks against source documents) to block or flag unsafe or low-quality responses. This is a defence-in-depth approach where each layer catches issues the others miss, ensuring the model behaves safely within its intended use case.
- ✗
Legal terms of service that constrain how developers can use Azure OpenAI
Why it's wrong here
Terms of service are commercial legal agreements between Microsoft and developers, setting contractual obligations and liability boundaries. They do not actively constrain AI outputs — they exist outside the inference pipeline. Guardrails are technical, real-time constraints embedded in the application layer (e.g., through Azure AI Content Safety APIs, prompt templates, or custom validation code), operating during each generation call, unlike static legal documents.
- ✗
Rate limits that prevent individual users from generating too many responses
Why it's wrong here
Rate limits are administrative or platform-level usage controls that cap the number of API requests or tokens per time window to protect backend resources and ensure fair use. They address throughput and demand, not the semantic quality or safety of AI responses. Guardrails instead evaluate the content being submitted and generated—rejecting or modifying harmful inputs/outputs—so a user could stay well within rate limits and still receive unsafe content without guardrails in place.
Go deeper
Related to this question
Learn chapter
Responsible AI Principles
Key term
Responsible AI
A framework of ethical principles and practices that ensure artificial intelligence systems are developed and deployed in a transparent, fair, accountable, and safe manner.
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.