AI-900 Practice Question: Describe Artificial Intelligence workloads and considerations

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Billing accountability — ensuring costs are tracked and charged to the correct Azure subscription

  Why it's wrong here

  Billing accountability, such as tracking Azure resource consumption and charging costs to the correct subscription, is a FinOps and cost governance practice, not responsible AI governance. Azure cost management tools, tags, and budgets help organizations control spending, but they say nothing about whether an AI system is biased, unsafe, or used without human oversight. AI accountability is concerned with the behavior and consequences of the model, including fair outcomes, safety, transparency, and the ability to explain and correct harms, rather than with monetary chargeback.

• ✓

  Humans remaining responsible for AI systems with oversight mechanisms and clear lines of accountability

  Why this is correct

  Accountability in Microsoft's responsible AI framework means that humans remain responsible for AI systems and can be held answerable for their outcomes, with oversight mechanisms and clear lines of ownership. Oversight includes human review processes, audit trails that record model behavior and decisions, continuous monitoring, and wherever necessary a human-in-the-loop or human-on-the-loop mechanism to override or stop an unsafe system. This principle is what transforms an automated model into a governable system by tying every significant AI-assisted decision to a specific person or team.

• ✗

  AI systems reporting their own mistakes and triggering automatic self-correction

  Why it's wrong here

  An AI system reporting its own mistakes or automatically retraining itself after detecting drift is a reliability or self-healing feature, not accountability. Accountability requires a moral agent — people — who can understand the system's purpose, accept legal and ethical consequences, and decide whether a correction was appropriate. Automated monitoring such as Azure Monitor or model retraining pipelines may help humans identify errors, but leaving evaluation and remediation to the system removes the human responsibility that the accountability principle explicitly requires.

• ✗

  Holding AI vendors legally accountable for damages caused by their models

  Why it's wrong here

  While vendor liability is an emerging legal and regulatory question, Microsoft's accountability principle in responsible AI places the duty on the people and organizations that deploy and operate AI systems, not on the model vendor alone. Deployers control use context, training data selection, deployment decisions, and monitoring, so they are positioned to perform impact assessments and maintain oversight. Vendor litigation may cover product defects, but it is not the operational accountability mechanism that Microsoft's framework requires — clear human ownership, audit trails, and the ability to intervene remain the core expectation.

Go deeper

Related to this question