Describe Artificial Intelligence workloads and considerations →mediumMultiple ChoiceObjective-mapped
AI-900 Practice Question: Describe Artificial Intelligence workloads and considerations
A corporation deploys an AI system that uses a deep neural network to recommend candidate profiles for job openings. The hiring managers cannot understand why a particular candidate was recommended or not. Which Microsoft responsible AI principle is most directly relevant?
⚠ Common exam trap
Candidates often confuse Transparency (explainability) with Accountability (who is responsible) or Fairness (bias), but the core issue is the inability to understand the model's reasoning, not who to blame or whether bias exists.
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
✓
Transparency
The scenario describes a deep neural network whose internal reasoning is opaque to users. Microsoft's Transparency principle requires AI systems to be interpretable and explainable, so that stakeholders can understand how decisions are made. This directly addresses the hiring managers' inability to see why a candidate was recommended or not.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Fairness
Why it's wrong here
Fairness is about detecting and mitigating bias, such as disparate impact or unequal error rates across demographic groups, often measured with parity metrics. The immediate problem here is not that the model discriminates against a protected class, but that stakeholders cannot understand why any particular recommendation was made. A system can be perfectly fair by statistical standards and still be a black box, so fairness does not address the observed deficiency.
- ✗
Reliability and safety
Why it's wrong here
Reliability and safety focus on whether the system performs consistently under expected conditions, degrades gracefully on out-of-distribution inputs, and includes safeguards against harmful failures. These concerns can be satisfied even when the model's internal reasoning is opaque, because they are validated through testing and monitoring rather than through explanation. The corporation's issue is about the absence of interpretable reasoning, not about the model crashing, misbehaving, or producing dangerous outcomes.
- ✓
Transparency
Why this is correct
Transparency requires that the behavior and rationale of an AI system be open to inspection, with decisions explainable in terms that stakeholders can understand and contest. A deep neural network's hidden layers make it difficult to trace how input features map to output predictions, which directly violates this principle by preventing users from knowing why a recommendation was generated. This principle is the specific stakeholder need in the scenario, as it underpins trust, auditability, and the ability to detect errors.
- ✗
Accountability
Why it's wrong here
Accountability is the principle that there must be a designated human or organization responsible for the AI system's outcomes, supported by governance, documentation, and remediation processes. In this scenario, the deficiency is not the absence of an accountable owner but the inability to trace a specific recommendation back to its reasoning, which is a transparency failure. While accountability ultimately depends on transparency to be meaningful, the immediate and direct violation described is the lack of explanation, not the lack of responsibility.
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
Transparency
Transparency in AI means that the inner workings, decision-making processes, and data used by an AI system are open, understandable, and auditable by humans.
About these practice questions
Courseiva writes every AI-900 question from scratch — 985 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. 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.