Guidelines for Responsible AI
Targeted AIF-C01 practice for this exam objective, with an explanation for every answer option.
- Certification
- AIF-C01
- Domain
- Guidelines for Responsible AI
- Questions
- 9
Practice AIF-C01 Guidelines for Responsible AI questions with exam-style scenarios, detailed explanations, and option-by-option rationale for this domain.
Example question from this objective
Each option is explained, so the reasoning behind the distractors is part of the answer.
A voice assistant performs well for the development team but poorly for users with regional accents. Which action best supports inclusive AI?
- Report only one aggregate accuracy across all users.
- Exclude affected accents from the published performance report.
- Test and improve it with consented, representative samples from the affected accents.
- Raise the model temperature for every supported accent.
Report only one aggregate accuracy across all users. — An aggregate can mask materially worse outcomes for a subgroup.
Exclude affected accents from the published performance report. — Excluding affected users hides rather than addresses the problem.
Test and improve it with consented, representative samples from the affected accents. — Representative evaluation and improvement across affected user groups expose and reduce exclusionary performance gaps.
Raise the model temperature for every supported accent. — Temperature changes output randomness and do not fix unrepresentative speech performance.