AIF-C01 Fundamentals of Generative AI Practice Question
Which TWO actions are best practices for reducing hallucinations in generative AI models? (Choose 2)
⚠ Common exam trap
AWS often tests the misconception that larger models or fine-tuning alone solve hallucinations, when in fact grounding techniques like RAG and explicit prompt constraints are the proven mitigations.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Use retrieval-augmented generation (RAG)
Option C, retrieval-augmented generation (RAG), is correct because it grounds the model's responses in externally retrieved, authoritative documents at inference time, so the model cites or conditions on real source content rather than relying solely on parametric memory, which measurably reduces fabricated facts. Option E, prompt engineering with clear instructions and constraints, is correct because explicit directives such as "answer only from the provided context," "say 'I don't know' if unsupported," and output-format constraints steer the model away from speculative generation and make hallucinations easier to detect. Option A, increasing model size, is not a reliable fix: larger models can be more fluent and confident while still hallucinating, and scale alone does not guarantee factual grounding. Option B, fine-tuning on proprietary data, mainly adapts style, format, and domain vocabulary and can even increase confident fabrication on facts not present in the tuning set, so it is not a primary hallucination-reduction best practice. Option D, using a smaller model, is also not a best practice for this goal, since reduced capacity generally lowers factual accuracy and reasoning rather than improving truthfulness.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Increase the model size
Why it's wrong here
Scaling parameters increases fluency and reasoning breadth, not factual grounding, so hallucinations remain unaddressed. It is tempting because larger models score higher on many benchmarks, and would be the right choice when the task demands richer general reasoning rather than verifiable, source-anchored answers.
- ✗
Fine-tune the model on proprietary data
Why it's wrong here
Fine-tuning on proprietary data teaches style, format and domain vocabulary, but it does not ground responses in verifiable sources, so unsupported claims persist. It is tempting because domain adaptation genuinely improves relevance, and would be the right choice when the goal is tone or task-specific output rather than factual accuracy.
- ✓
Use retrieval-augmented generation (RAG)
Why this is correct
Retrieval-augmented generation grounds responses in documents fetched at inference time, so the model conditions on supplied evidence rather than relying solely on parametric memory. This directly addresses the hallucination constraint in the stem by anchoring output to verifiable source content, reducing fabricated claims when the retrieved passages are relevant and accurate.
- ✗
Use a smaller model to limit complexity
Why it's wrong here
Model size does not govern factual grounding; a smaller model can hallucinate as readily as a larger one, so this does not address the cause. It is tempting because smaller models reduce latency and cost, and would be the correct choice when those constraints, not accuracy, drive the design.
- ✓
Apply prompt engineering with clear instructions and constraints
Why this is correct
Clear instructions and explicit constraints narrow the model's output distribution, reducing the chance it fabricates unsupported content. This directly satisfies the stem's goal of reducing hallucinations by grounding generation in specified boundaries, formats and scope, rather than leaving the model to infer intent from ambiguous prompts.
Go deeper
Related to this question
About these practice questions
This AIF-C01 question is part of Courseiva's 862-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AIF-C01 practice question is part of Courseiva's free Amazon Web Services certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AIF-C01 exam.