Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Module: Week 5 of 16 · Fall 2026 · in-person, two 75-minute sessions
Objective covered: Objective 2 — Use effective prompting techniques — including conversation, content, emphasis, meta-prompting, structured prompts, examples, and simulations — to produce high-quality, well-verified AI output.

This file holds two pieces: (A) the Module 5 Overview page ("Start Here") and (B) the Week 5 Announcement that drips out when the module opens. Dates assume a Tuesday/Thursday lecture pattern with Week 5 meeting Tue Sep 30 and Thu Oct 2, and end-of-week work due Sunday Oct 4, 11:59 p.m. Adjust day-of-week and times to match your section.

(A) Module 5 Overview — Start Here

Welcome to Week 5: Show the AI What You Mean

This is your home base for the week. Read it first, then work the checklist below from top to bottom.

You've now built a conversation (Week 3), provided content and used emphasis (Week 3), meta-prompted your way to a structured template (Week 4), and named every component of a great prompt (Week 4). But here's a gap you've probably already hit: sometimes you know exactly what you want, you've described it in every way you can think of — and the AI still misses the voice, the format, or the level of detail. The fix is almost always an example.

This week we complete Objective 2 by adding the most powerful and underused prompting technique: showing the AI what you want instead of just telling it. We'll cover the vocabulary (zero-, one-, and few-shot prompting), use examples to teach voice, format, and how to handle sensitive information, and build out a full control toolkit: specifying structure, setting constraints, naming a count, regenerating (and why it doesn't fix facts), asking for expansion, requesting sources (and verifying them), and asking for guidance.

The week's big question

"When should you show the AI what you want — and how do you keep it from over-generalizing from your examples?"

By Sunday you'll be able to label any prompt as zero-, one-, or few-shot; build a few-shot prompt that teaches a specific voice or format; apply at least four control moves to sharpen AI output; and know exactly what to do when the AI provides sources (verify every one).

By the end of this week, you can…

Use this as a checklist. If you can do all five out loud, you're ready for the quiz.

• [ ] Define zero-shot, one-shot, and few-shot — and explain that few-shot means a few examples, not exactly one.
• [ ] Build a few-shot prompt that teaches a specific voice, format, or transformation (e.g., PII removal) using two to five examples.
• [ ] Apply control moves: specify a count, request a structure, set a constraint, ask for expansion, and regenerate — and explain the difference between regenerating and fixing.
• [ ] Request sources from an AI and explain why you must verify every link or citation before trusting it.
• [ ] Catch drift or over-generalization when the AI over-applies patterns from your examples and describe the fix.

What's due this week, and when

Work these in order — each one gets you ready for the next.

Reminder: you are required to use AI on the tutorial, discussion, assignment, practice, and Studio — that's the whole point. AI is not allowed on Quiz 5, which checks that you understand. And every week you catch the AI's mistakes — this week it's catching drift from your examples and fabricated sources.

Late policy reminder: 10% off per day late.

How to succeed this week

• Start with two examples, not zero. Before you try to describe a format or voice in words, write out two good examples of the thing you want. Paste them in and say "continue this pattern." You'll often get better results faster than any description would give you.
• Count your shots. When you build a few-shot prompt, you're giving the AI a pattern to generalize from — two to five examples is the sweet spot. One example is one-shot; several is few-shot; "few" never means exactly one.
• Verify every source. This cannot be over-stated: when you ask an AI for citations, links, or references, open every single one before trusting it. Regenerating a bad citation gives you a different bad citation. The fix is manual verification.
• Catch the drift. When you give examples, read the output specifically for over-generalization: did the AI turn a stylistic quirk into a rule? Did it mechanically repeat a word or structure that only appeared once in your examples? That's drift — and you fix it by adding another example that shows the range or by adding an explicit constraint.
• Asking for guidance is a control technique too. "What would help you do this better?" or "What am I not thinking about?" is one of the most underused moves in prompting.

You're building a real toolkit this week — not theory. By the Studio, you'll have a specific format or voice the AI can reproduce on demand. See you Tuesday.

(B) Welcome Announcement — Module 5

Release setting: post on the module's start day (offset = 28 days from module 1), i.e., Tue Sep 30, 2026 — not before. If your platform won't preserve the scheduled date on import, post this as a draft labeled "Release: Tue Sep 30."

Subject: Week 5 — the most underused prompting move (and it's not what you think)

Hi everyone,

Here's a quick challenge: think of a task you've given an AI where the output was almost right but the voice was slightly off, or the format wasn't quite what you pictured, or the level of detail was close but not exactly there. What did you do? You probably rewrote the prompt, described what you wanted more carefully, or just tried to live with the result.

This week, we fix that problem for good — with examples.

Week 5 — Prompting III: Examples, Structure & Control is where the prompting arc completes. We add the technique that makes everything else sharper: showing the AI what you want instead of just describing it. We'll cover the zero/one/few-shot vocabulary, use examples to teach a specific voice or format, handle sensitive information safely, and build a full control toolkit — specifying structure, count, constraints, regenerating intelligently, asking for expansion, requesting sources, and asking for guidance.

Three things not to miss:
1. The few-shot/one-shot distinction — the quiz specifically tests this. "Few-shot" means a few examples (typically two to five). One example is "one-shot." They're not the same. Memorize that now.
2. Requesting sources ≠ verified sources — every citation the AI gives you must be verified before you use it. Regenerating a bad citation gives you a different bad citation. Verification is always manual.
3. Studio 5 ("Few-Shot Your Format/Voice") is the hands-on build: give the AI 2–3 examples to teach it a specific format or voice, then catch where it drifts or over-generalizes. Start it early — it takes real iteration.

A heads-up on how AI works in this course:
1. You are required to use AI on the tutorials, discussions, assignments, practice, and the weekly Studio.
2. AI is not allowed on Quiz 5, which checks that you understand the concepts.
3. Every week we catch the AI's mistakes — this week, you'll catch drift from your examples and fabricated citations.

Quiz 5, Studio 5, Discussion 5, and Assignment 5 all close Sunday Oct 4. Initial discussion post is due Friday Oct 2.

See you Tuesday with your examples ready.

Prof. Quinn