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.

• ✗

  Transparency

  Why it's wrong here

  Transparency is about making an AI system's behavior, limitations, and data usage understandable to its users and stakeholders, for example through model documentation, interpretability tools, or clear labeling of AI-generated content. While transparency is a prerequisite for oversight, it does not by itself create a formal mechanism for an individual to contest a decision or receive a remedy. Accountability, however, explicitly requires such a mechanism—clear human responsibility, audit trails, and channels for appeal—so choosing transparency would fail to address the governance and redress aspects of the scenario.

• ✓

  Accountability

  Why this is correct

  Accountability is the responsible AI principle that requires organizations to take ownership of AI systems' outcomes, assign clear human responsibility, and implement processes through which people can contest decisions or seek redress. This includes human oversight of consequential automated decisions, ongoing auditing and impact assessments, and documented escalation paths for affected users. In practice, an accountable system must have a 'human in the loop' who can override or review the AI's output, as well as a formal appeal channel—this directly matches the described ability to raise concerns and seek redress.

• ✗

  Reliability

  Why it's wrong here

  Reliability and safety focus on the technical consistency and robustness of an AI system, ensuring it performs as expected under normal and unusual conditions, handles edge cases gracefully, and does not cause accidental harm. A reliable system might be extremely consistent yet still lack any formal channel for users to ask questions or contest its output if things go wrong. Accountability is a broader governance concern that includes oversight, audit trails, and appeal processes; therefore, reliability alone does not satisfy the requirement that humans be responsible and decisions be challengeable.

• ✗

  Fairness

  Why it's wrong here

  Fairness is concerned with preventing and mitigating bias so that AI systems treat individuals and groups equitably, typically measured through bias metrics and evaluated across protected attributes such as race, gender, or age. While fairness is an important outcome-related property, it is not a mechanism for assigning responsibility or enabling contestation of a decision. Accountability, by contrast, is the governance principle that ensures humans remain in charge and that there are clear procedures for raising concerns and seeking redress, which is distinct from the statistical and design work involved in making a system fair.

Go deeper

Related to this question