AI-900 Practice Question: Describe features of generative AI workloads on Azure
What is 'hallucination' in large language models and what techniques help reduce it?
⚠ Common exam trap
Many exam-takers confuse hallucination with other common AI issues like modality switching (A), anthropomorphism (C), or data memorization (D), because all involve unexpected or problematic model behavior, but only B captures the core definition of generating confident falsehoods.
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
✓
When a model generates confident but factually incorrect or fabricated information
Hallucination in large language models (LLMs) refers to the generation of text that is confident, coherent, and plausible-sounding but factually incorrect or entirely fabricated. This occurs because LLMs are probabilistic next-token predictors trained on vast datasets, not databases of verified facts; they lack a built-in mechanism to distinguish truth from fiction. Techniques to reduce hallucination include grounding outputs with retrieval-augmented generation (RAG) using Azure AI Search, prompt engineering with system messages that constrain responses to verified sources, and fine-tuning with human feedback (RLHF) to penalize factual errors.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
When a model generates images instead of text in response to a text prompt
Why it's wrong here
Unexpected output modalities stem from model configuration or task mismatch, not hallucination. Hallucination is confined to the textual domain: an LLM fabricates plausible-sounding but factually wrong statements while predicting tokens. Producing images instead of text reflects routing, decoding, or multimodal integration issues, not an error of factual content. Therefore this option is incorrect.
- ✓
When a model generates confident but factually incorrect or fabricated information
Why this is correct
This is correct because hallucination in LLMs is confident confabulation: the model generates text that is grammatically fluent and internally consistent yet factually false. Autoregressive token prediction selects the most probable next token from learned distributions, with no intrinsic verification against external knowledge, so the output can state fabricated events, citations, or numbers as if true. The 'confidence' is stylistic, not epistemic — nothing in the architecture knows whether the claim is grounded in reality.
- ✗
When users imagine the AI is sentient due to very convincing responses
Why it's wrong here
This describes the ELIZA effect or user anthropomorphism, a human perception phenomenon, not a model failure mode. Hallucination is an objective technical property of generated text: the output contradicts verifiable facts, regardless of whether any user feels the model is sentient. A convincing, human-like response can still be factually correct; conversely, a hallucinated answer rarely depends on the user's subjective impression of sentience. Thus it is wrong.
- ✗
When a model's training data contains copyrighted material it memorises
Why it's wrong here
Memorisation is a distinct failure mode involving data leakage or replication, where the model reproduces training sequences verbatim because it has overfitted or stored them. Hallucination, by contrast, generates novel false information that never appeared in any training source — it invents rather than recalls. Copyright infringement concerns are about fidelity to original content, while hallucination is about confabulated inaccuracy, so this option is incorrect.
Go deeper
Related to this question
Learn chapter
Azure Machine Learning Studio
Key term
Azure AI Search
Azure AI Search is a cloud search service that uses artificial intelligence to index and search through large amounts of data, returning relevant results that feel as smart as a Google search.
Key term
RAG
Retrieval-Augmented Generation (RAG) is a technique that combines information retrieval with text generation to produce more accurate and contextually relevant answers from large language models.
About these practice questions
Courseiva writes every AI-900 question from scratch — 985 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or 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 AI-900 practice question is part of Courseiva's free Microsoft 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 AI-900 exam.