AI and Hallucinations Still: Why Even GPT-5 Lies

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-builders-foundations-AI-and-hallucinations-still-r13a9-teen

1. What is an AI hallucination?

   1. When an AI gives different answers to the same question each time
   2. When an AI takes too long to process a request
   3. When an AI generates false or made-up information and presents it as true
   4. When an AI refuses to answer a question because it lacks data

2. According to the 2026 information in this topic, what happened to hallucination rates?

   1. They dropped but did not reach zero
   2. They disappeared completely
   3. They increased significantly
   4. They stayed exactly the same as in 2023

3. What is the recommended solution for dealing with AI hallucinations?

   1. Only using AI for simple, uncontroversial topics
   2. Developing verification habits to check AI outputs
   3. Waiting for developers to create perfect AI models
   4. Avoiding AI tools entirely

4. Which type of information is an AI most likely to fabricate?

   1. Basic math problems
   2. Very simple and commonly known facts
   3. General concepts that most people understand
   4. Specific book titles, court cases, and historical dates

5. What does the phrase 'AI lies most confidently when it is most wrong' mean?

   1. The more incorrect AI's response is, the more certain it sounds
   2. AI can detect when it is wrong and admits it
   3. AI only lies when it has been deliberately programmed to deceive
   4. AI is equally confident about all its answers

6. What is 'grounding' in the context of AI?

   1. The process of training AI on more data
   2. A method to make AI respond faster
   3. A technique to reduce hallucinations by connecting AI to external information sources
   4. A way to make AI sound more human

7. What does RAG stand for and what does it do?

   1. Random Answer Generation — it makes AI choose answers randomly
   2. Retrieval-Augmented Generation — it reduces hallucinations by fetching verified information
   3. Reasonable Audio Generation — it creates spoken responses
   4. Rapid AI Growth — it helps AI learn faster

8. Why is checking specific details like book titles especially important?

   1. Because books are not important to verify
   2. Because AI cannot talk about books
   3. Because books are always mentioned correctly by AI
   4. Because AI is most likely to make up specific, verifiable details

9. What should you do if an AI's answer sounds too specific to be true?

   1. Accept it as likely true since it sounds confident
   2. Ignore it and ask a different question
   3. Share it on social media immediately
   4. Check the information using another source

10. Why might searching a claim on Wikipedia be a good verification step?

    1. Because it allows quick checking of factual claims against a generally reliable source
    2. Because Wikipedia is owned by AI companies
    3. Because AI cannot lie about information found on Wikipedia
    4. Because Wikipedia is always 100% accurate

11. What does it mean that AI presents fabricated information with 'total confidence'?

    1. AI always knows when it is lying
    2. AI only lies when it is unsure
    3. AI presents false information using the same confident tone as true information
    4. AI can accurately measure its own certainty

12. What habit does the lesson say 'separates AI users from AI victims'?

    1. Never questioning AI responses
    2. Only using AI for fun tasks
    3. Believing everything AI says without question
    4. Verifying AI outputs before accepting them

13. Why might asking an AI to cite specific book chapters be a useful test?

    1. Because it creates an opportunity to verify specific claims
    2. Because AI always knows exact chapter information
    3. Because this is the only safe way to use AI
    4. Because book chapters cannot be fabricated

14. What is a hallucination rate?

    1. The number of users who report AI errors
    2. The speed at which an AI generates responses
    3. The amount of data an AI was trained on
    4. The percentage of time an AI produces incorrect or fabricated information

15. Which tool mentioned in the lesson is specifically designed to show sources for its answers?

    1. Perplexity
    2. Claude
    3. Gemini
    4. ChatGPT