AI EMAIL COURSE - KNOWLEDGE CHECK QUIZ

QUIZ METADATA:
→ Course module: "Foundations of Prompt Engineering"
→ Audience: Beginners (no prior AI experience assumed)
→ Format: 8 questions, 5-7 min completion
→ Passing score: 70 percent (6/8 correct)
→ Outcome on pass: Unlock module certificate + next module
→ Outcome on fail: Show remedial resource, allow 1 retake

QUESTIONS & SCORING:

Q1: MULTIPLE CHOICE (1 point)
"What is 'temperature' in a language model?"
A) The number of words in a response
B) How random or deterministic the model's output is
C) The cost per API call
D) The version number of the model
→ Correct answer: B
→ Explanation (on wrong): "Temperature controls randomness. High temperature (0.8+) = creative, varied outputs. Low (0.0-0.3) = predictable, consistent."

Q2: MULTIPLE CHOICE (1 point)
"Which prompt technique is best for multi-step problems?"
A) Few-shot examples
B) Chain-of-thought prompting
C) Temperature tuning
D) Token limiting
→ Correct answer: B
→ Explanation: "Chain-of-thought breaks complex tasks into steps, improving accuracy on reasoning problems."

Q3: FREE TEXT (1 point) - Manual grading or keyword match
"Write a 1-sentence prompt that asks AI to explain quantum computing like a 10-year-old."
→ Sample acceptable answer: "Explain quantum computing in simple words for a child."
→ Rubric: Award 1 point if it includes (1) AI instruction, (2) topic, (3) audience level

Q4: MULTIPLE CHOICE (1 point)
"What is a 'hallucination' in AI?"
A) A wrong answer that sounds confident
B) An error in the training data
C) A limitation of old models only
D) Rare and not a real concern
→ Correct answer: A
→ Explanation: "Hallucinations are false outputs presented confidently. Always verify AI outputs against trusted sources."

Q5: MATCHING (2 points - 1 per match)
Match prompt technique to use case:
→ Few-shot examples ← When you need a specific output format
→ System prompt ← Define the model's personality/role
→ Prompt injection ← Don't use this (security risk)
→ Temperature tuning ← Adjust randomness of creative tasks

Q6: TRUE/FALSE (1 point)
"AI models understand language the way humans do."
→ Answer: FALSE
→ Explanation: "Models recognize patterns in text but don't have human understanding or consciousness."

Q7: MULTIPLE CHOICE (1 point)
"You ask AI for code and it returns a deprecated Python library. What should you do?"
A) Use it as-is
B) Ask AI to fix it, verify against current docs, test thoroughly
C) Assume AI always has latest info
D) Use ChatGPT instead
→ Correct answer: B

Q8: SHORT ESSAY (1 point)
"Describe one limitation of AI models and how to work around it."
→ Rubric: Award 1 point if answer identifies (1) real limitation (hallucination, bias, knowledge cutoff, lack of reasoning), (2) practical workaround (prompt technique, human review, chaining, etc)

SCORING & RESULTS PAGE:

If score >= 70 percent:
→ Display: "Congratulations! You've mastered prompt basics."
→ Badge: Unlocked "Prompt Engineer Foundation" certificate
→ Next: Button to access Module 2
→ Email: Congratulations message with certificate PDF

If score < 70 percent:
→ Display: "Great effort! Review these topics:"
→ List: Missed question topics (e.g., "Prompting Techniques")
→ Resource: Link to lesson on that topic (video + article)
→ Retry: "Retake Quiz" button (allow 1 attempt, 24hr wait)
→ Encouragement: "Most people pass on retake!"