2.2Standard
Build a chatbot for internal policy questions.
Do not return unsupported assertions.
Choose two appropriate controls.
× Incorrect
Ground answers in trusted documents with citations and provide a safe unknown or human-review path when evidence is missing.
Detailed explanation
CorrectState that the answer is unknown and route to a person when evidence is missing
This avoids unsupported certainty.
IncorrectTreat fluent prose as factual
Plausibility does not prove correctness.
IncorrectTell the model to present guesses as facts
Hiding uncertainty encourages misuse.
CorrectRetrieve trusted documents and answer with sources
External evidence improves verifiability and grounding.
Check official documentation
Compare the following points with the official documentation.
Amazon Bedrock Knowledge Bases・Guardrails公式資料とAIF-C01 Domain 2.2を確認する。
Official references
What to confirm
流暢さと事実性を区別し、接地と安全なフォールバックを説明できる。
Key points
- Grounding
- Sources
- Fallback
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Hallucination
A plausible-sounding model output that is unsupported or factually incorrect.
Grounding
Connecting an output to trusted external data, retrieved documents, or verifiable evidence.
Question IDAWS-AIF-116
EnvironmentAWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
Last technical review2026-08-20
Report an error or rights issue