Week 1 · Lecture outline

Week 1 — Lecture Outline · Welcome to the AI Revolution

Using Artificial Intelligence · AI 101 Fall 2026 · Prof. Quinn Fictional sample

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Objective covered: Objective 1 — Explain what generative AI and large language models are, how they work conceptually, and their core capabilities and limits — including why AI can be "confidently wrong."
SLOs touched: A (produce high-quality results with AI through strong prompting) · B (evaluate and use AI critically)
Meeting pattern: 2 sessions × 75 min = 150 min. Segment minutes below total ~150; scale to your own pattern.

Week at a Glance


The week's big question	"What is this thing I'm talking to — and how do I have to think to get the most out of it?"
By the end of the week, students can…	(1) define generative AI and use AI / generative AI / LLM / AGI correctly; (2) adopt the general→specific, iterative mindset and the "the machine has no brain" stance; (3) set up an approved assistant and run a useful first conversation; (4) explain the "what if…" paradigm and why the human stays the judge.
Key vocabulary	artificial intelligence (AI), generative AI, large language model (LLM) / "the model," chatbot/assistant, prompt, AGI (artificial general intelligence), context, iteration, hallucination (preview), prompt engineering (preview)
Materials	slides (Deck 1), the week's readings + video links, one approved assistant (ChatGPT / Claude / Gemini / Copilot) for the live demos, the AI-critique moment, and the tutorial
Timing note	8 segments, ~150 min total. Session 1 = Segments 1–4 (~75). Session 2 = Segments 5–8 (~75).

Segment 1 — Hook & the Promise (8 min) · Session 1 opens

Hook. Put one line on a slide and make the room vote: "When a chatbot answers you, is it thinking?" Take a show of hands. Then read a short, fluent, wrong AI answer to a question in your field (have one ready — e.g., a confidently invented "fact" about your campus). "It sounds like it understands. It's smooth, it's confident, and part of it is made up. So whatever it's doing, it isn't 'knowing.' Today we figure out what it is doing — because if you get that wrong, you'll either over-trust it or give up on it."

The promise (write it on the board): "By Friday you'll be able to say exactly what generative AI is, use the right words for it, run a conversation that actually saves you time, and know the one habit that separates people who get great results from people who rage-quit."

Why it matters line (memory hook): "The machine has no brain — so you bring yours."

Segment 2 — What Generative AI Actually Is (20 min)

Plain language first. Generative AI is software that creates new content — text, images, audio, code — in response to a request. The chatbots you've used (ChatGPT, Claude, Gemini, Copilot) are powered by a large language model (LLM), which people in the field just call "the model." At a plain level, an LLM works by predicting the next chunk of text ("token") over and over, based on patterns it absorbed from an enormous amount of training text. It is astonishingly good at producing fluent, relevant language — and it has no understanding, no beliefs, and no memory of you beyond the current conversation. (We go one level deeper on "how" next week; today it's the mental model.)

Get the vocabulary straight (put this on one slide):
- Artificial intelligence (AI) — the broad field of getting computers to do things that used to need human intelligence. The umbrella term.
- Generative AI — the slice of AI that generates new content. This course lives here.
- Large language model (LLM) / "the model" — the text-prediction engine inside a chatbot. The chatbot is the app; the model is the engine.
- AGI (artificial general intelligence) — a hypothetical future AI that could do any intellectual task a human can. It does not exist today — today's tools are powerful but narrow. Don't confuse the two.

Memory hook: "AI is the field, generative AI makes things, the LLM is the engine, AGI is the sci-fi goal."

The clarification students always need: the model is not looking things up in a database and it is not a search engine. It's generating likely text. That's why it can be fluent and wrong at the same time — a property we'll exploit (for creativity) and guard against (for facts) all term.

Segment 3 — The Mindset That Makes AI Useful (22 min)

Plain language first. The single biggest predictor of whether someone gets value from AI isn't the tool — it's their mindset. Two habits do most of the work.

Habit 1 — General → specific. Don't agonize over the "perfect" opening prompt. Start with a clear-enough request, see what comes back, then steer: "shorter," "for a non-expert," "give me three options," "now in a table." You're a director giving notes, not a person entering a one-shot search.

Habit 2 — Iterate; it's a conversation. The first answer is a draft, not a verdict. The magic is in the back-and-forth. People who treat the chat box like Google ("type once, judge the result, leave") get mediocre results; people who treat it like a conversation with a fast, eager intern get great ones.

The stance underneath both — "the machine has no brain." Because the model predicts plausible text rather than knowing things, you stay the brain: you set the goal, supply the judgment, and check the facts. The AI is the most capable intern you've ever had — tireless, fast, widely read, occasionally confidently wrong, and in need of direction.

One fully worked example (do it live on the projector).

Weak first prompt: "Give me tips for studying." → generic listicle.
Steer once: "I'm a nursing major studying for an anatomy exam in 5 days. Give me a day-by-day plan." → much better.
Steer again: "Put it in a table, and add one active-recall technique per day." → genuinely useful.
Point made: the quality jumped because the human iterated — not because the model "tried harder."

Memory hook: "Start general, then get specific — and keep talking."

Segment 4 — Misconceptions + Quick Interaction (22 min) · Session 1 closes (~75)

Name the misconceptions out loud, then cure each:

❌ "The AI understands me / it's thinking."
✅ Cure: it predicts likely text from patterns; there's no understanding or belief behind it. It can model understanding well enough to be useful — but treat fluency as a style, not a guarantee of truth.
❌ "A chatbot is just a fancy search engine."
✅ Cure: a search engine finds existing pages; a generative model writes new text that may or may not be accurate. Different tool, different failure modes. (More in Week 2.)
❌ "If the first answer is bad, the AI is useless."
✅ Cure: the first answer is a draft. Value comes from steering it — general → specific, with follow-ups.
❌ "AGI is basically here / ChatGPT is an AGI."
✅ Cure: today's tools are powerful but narrow; AGI does not exist. Useful to know so you neither over-trust nor dismiss what's real today.

Interaction — Think-Pair-Share (rapid-fire, ~10 min):
Put four statements on a slide; for each, students decide true or false, and why, solo (30 sec), compare with a neighbor (1 min), then vote. Suggested items: "The LLM is the app you open." (F — the model is the engine inside it) · "Generative AI can create images, not just text." (T) · "If you give the same prompt twice you always get the same answer." (F — outputs vary) · "AGI is what powers today's chatbots." (F). Have them name the deciding idea each time.

Segment 5 — The "What If…" Paradigm & Adaptability (20 min) · Session 2 opens

Hook back in: "Last session: what AI is. Today: why it changes how you work — and how to keep up as it changes."

Plain language first — AI lowers the cost of trying things. Before, testing an idea ("what would a cover letter for this job look like?" "how might I outline this essay three different ways?" "what could go wrong with this plan?") cost real time, so you tried one. Generative AI makes a first draft nearly free, so you can ask "what if…" ten times and pick the best. The skill shifts from producing a first draft to directing, comparing, and judging lots of drafts.

Worked example (live): take a real task from a student ("email asking a professor for a recommendation"). Generate three versions — formal, warm, brief. Compare out loud. The point: the value wasn't the AI's writing; it was the cheap exploration plus your judgment about which fits.

Adaptability — the meta-skill. These tools change monthly. The durable skills are not "which button is where" but how to think: prompt well, verify, pick the right tool, and stay willing to relearn. "Don't memorize the menu; learn the moves." We'll point at official docs all term precisely because the specifics move.

Memory hook: "AI makes 'what if?' cheap — so try more, and judge well."

Segment 6 — Live Demo: Your First Real Conversation (20 min)

Set it up: "Watch me set up and run a useful conversation from scratch — this is exactly the move I want you doing in the tutorial and Studio this week."

Do it live (narrate every step):
1. Open an approved assistant (ChatGPT, Claude, Gemini, or Copilot — name that any free account works).
2. State a real goal with context: "I'm a first-year business major. Help me plan a 5-day study schedule for my first marketing quiz. Ask me 2 questions first if you need to." — model giving context and inviting clarifying questions.
3. Iterate: make it shorter; ask for a table; ask "what am I missing?"
4. Ask for guidance: "What's a good way to use you to study that I might not have thought of?" — model treating the AI as a coach, not just an answer machine.
5. The verify beat (preview): ask it a checkable fact relevant to the task and say aloud, "I'd confirm this before trusting it."

Land the key idea: a good first conversation = context + a clear goal + iteration + your judgment. Nobody is born knowing this; it's a learnable move, and you'll practice it all week.

Misconception + cure:
- ❌ "Good prompters just know the magic words."
✅ Cure: there are no magic words. They give context, iterate, and verify. That's the whole trick — and it's teachable.

Segment 7 — Setting Up & Privacy First Touch (20 min)

Part A — get set up (do this together):
- Everyone opens an approved assistant and confirms they can send a message. Troubleshoot logins now so nobody is blocked on the tutorial.
- Note the free tiers are enough for this week; Claude Cowork desktop comes later (Week 11) for the automation weeks.

Part B — a first, plain word on privacy (full unit is Week 15):
- What you type into a consumer AI tool may be stored, and on some tools/settings may be used to improve the model. So from day one: don't paste anything you wouldn't be okay seeing made public — no passwords, no other people's private data, no confidential work material. "Would I be fine with this on a billboard?" If not, don't paste it.
- This isn't fear-mongering; it's the same caution you'd use with any new online service. We'll get specific (ToS, anonymizing, enterprise controls) in Week 15.

Memory hook: "If you wouldn't put it on a billboard, don't paste it into a free AI tool."

Segment 8 — Technology Workflow + AI-Critique, Callback & Hand-off (18 min) · Session 2 closes (~75)

Technology workflow — the first-conversation habit, on demand:
1. State your goal and give context (who you are, what it's for).
2. Invite clarifying questions ("ask me anything you need first").
3. Iterate — general → specific, with follow-ups.
4. Judge and verify — you decide what's good, and you check anything factual.

AI-critique moment (students verify, not consume) — the course through-line begins here:

Ask an approved assistant: "Give me five interesting facts about Silver Oak University." (Use your own campus.) The model will likely produce confident, specific, partly invented "facts."
Now be the judge: which of these could you actually verify? Which sound made up? This is the headline lesson of the whole course — fluency is not truth. You'll catch exactly this kind of confident fabrication in this week's Studio.

Callback + tease:
- Callback: "Everything this term rides on this week — a correct mental model of what AI is, the general→specific iterative mindset, and the habit of staying the judge."
- Tease next week: "We said the model 'predicts the next chunk of text.' Next week we open the hood — how it does that, what a token and a context window are, and why it's so confidently wrong sometimes."

Hand-off (the week's graded work):
- Lecture Tutorial 1 (AI tutor, share-link submission) — the terminology and the mindset.
- Quiz 1 (no AI), Discussion 1 ("Is It Thinking? / Spot the Bad Mental Model"), and Assignment 1 (define the terms; fix weak prompts; explain the mindset).
- AI Build Studio 1 — "Your First Great Prompt" — refine a weak prompt into a strong one, then catch where the AI's draft invents or overreaches.

Instructor FAQ — Common Stumbles

Student says / does	Quick cure
"So the AI is basically Google."	Search finds existing pages; generative AI writes new text that can be wrong. Different tool, different failure modes.
Confuses LLM and chatbot.	The chatbot is the app; the LLM ("the model") is the prediction engine inside it.
Thinks AGI is here.	Today's tools are powerful but narrow; AGI doesn't exist yet. Useful so you neither over-trust nor dismiss real capability.
"It gave me a bad answer, so it's useless."	First answer = draft. Steer it: general → specific, with follow-ups.
"It understands me."	It predicts plausible text; no understanding behind it. Treat fluency as style, not proof of truth.
Pastes private/confidential info to test it.	"Would I put this on a billboard?" If not, don't paste it. (Full privacy unit: Week 15.)
Hunts for "magic words."	No magic words — context + iteration + verification. That's the skill, and it's teachable.

Scope flag

This outline stays within Objective 1 at the mental-model level (what genAI is; the vocabulary; the mindset; the "what if" paradigm). The deeper "how it works" mechanics — tokens, the context window, training, and why hallucination happens — are Week 2 and only previewed here. Real products (ChatGPT, Claude, Gemini, Copilot) are named factually; the instructor and institution are fictional. The buyer-facing verb for the product is "generates."

~ Prof. Quinn's edition · Fall 2026 · built with thecoursemaker.com