AWS Certified AI Practitioner (AIF-C01) — Question 301

An AI practitioner notices a large language model (LLM) is generating different responses for the same input across multiple invocations.

Which risk of AI does this describe?

Answer options

• A. Hallucinations
• B. Nondeterminism
• C. Accuracy
• D. Multimodality

Correct answer: B

Explanation

Nondeterminism is the characteristic of an AI system where the same input can yield different outputs across different executions due to probabilistic sampling methods. Hallucinations refer to the generation of factually incorrect or nonsensical information, whereas accuracy measures the correctness of the output. Multimodality is the ability to handle multiple types of inputs and outputs, not a risk of inconsistent performance.