Exams / AIF-C01 / Securing AI systems / Question 5
AIF-C01 practice question: Securing AI systems (5 of 5)
Security, Compliance, and Governance for AI Solutions A free question from our original AIF-C01 practice exam. Try it, then open the answer to see why each option is right or wrong.
A pension provider's RAG assistant sometimes adds details that are not in the retrieved policy documents. The provider wants to automatically detect and block responses that are not supported by the source material. Which Amazon Bedrock Guardrails feature fits?
Choose one answer.
- AA. Denied topics
- BB. Word filters
- CC. Content filters for hate and violence
- DD. Contextual grounding checks
Show the answer and explanation
Answer: D
- A. Wrong: Denied topics block whole subjects. They do not check whether an answer is supported by the source.
- B. Wrong: Word filters block listed words or phrases, not unsupported claims.
- C. Wrong: Content filters target harmful categories. An unsupported pension detail is not harmful content in that sense.
- D. Right: Correct. These detect responses that are not grounded in the source information or are irrelevant to the query.
Explanation
Contextual grounding checks in Amazon Bedrock Guardrails compare a response with the reference source and the user's query, and flag answers that are ungrounded or irrelevant. They help detect hallucinations in RAG applications. Automated Reasoning checks are another option, validating responses against defined logical rules.
Study
Keep practising
Answer questions with a score at the end, or practise just this topic with the explanation after each one. No sign-up needed.