AWS Certified AI Practitioner (AIF-C01) — Question 196

An ecommerce company is using a chatbot to automate the customer order submission process. The chatbot is powered by AI and is available to customers directly from the company's website 24 hours a day, 7 days a week.

Which option is an AI system input vulnerability that the company needs to resolve before the chatbot is made available?

Answer options

• A. Data leakage
• B. Prompt injection
• C. Large language model (LLM) hallucinations
• D. Concept drift

Correct answer: B

Explanation

The correct answer is B, prompt injection, which refers to a method where malicious inputs can manipulate the AI's responses. Data leakage (A) pertains to unauthorized data exposure, LLM hallucinations (C) involve the model generating false information, and concept drift (D) describes changes in data patterns over time, none of which directly address input vulnerabilities like prompt injection does.