Exams / AIF-C01 / GenAI capabilities, limits and value / Question 1
AIF-C01 practice question: GenAI capabilities, limits and value (1 of 2)
Fundamentals of Generative AI 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 travel firm's chatbot confidently tells a customer about a hotel spa that does not exist. The answer sounds fluent and plausible. What is this limitation of generative AI called?
Choose one answer.
- AA. Overfitting
- BB. Latency
- CC. Data drift
- DD. Hallucination
Show the answer and explanation
Answer: D
- A. Wrong: Overfitting is a training problem where a model memorises its training data. It does not describe a confident, invented answer.
- B. Wrong: Latency is how long a response takes. The problem here is the content, not the speed.
- C. Wrong: Data drift is a change in production data over time. It does not describe a single invented fact.
- D. Right: Correct. A hallucination is output that sounds plausible but is false or not supported by any source.
Explanation
Generative models predict likely text, so they can produce fluent statements that are not true. These hallucinations are one of the main drawbacks of generative AI, together with nondeterminism and limited interpretability. Grounding answers in trusted data with RAG and checking responses with guardrails reduce the risk.
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.