Week 3 — Discussion (Adaptive Learning) · "Where's the Line? / Sycophancy in the Wild"

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Objective: Objective 2 (prompting: conversation, content-provision, sycophancy) · SLO B (reason critically about how to use AI responsibly)
This is Discussion 3 of 15 · Discussions group = 10% of the grade · Worth 20 points
Format: adaptive learning — instead of writing a post cold, you'll think it through in a real-time dialogue with your own AI, then post the short summary the AI writes with you (plus a link to your chat).

Part 1 — Student Instructions (read this first)

What this is. You'll reason through a genuinely contested question — where is the line for pasting your own (or someone else's) content into a free AI tool? — and then analyze a sycophantic exchange — diagnosing what went wrong when an AI agreed with everything its user said. You'll do both in a back-and-forth conversation with an AI assistant. The AI draws out and challenges your thinking; it does not hand you the answer. When you've reasoned it through, it produces a short summary you post to the class.

How to run it (about 15–20 minutes):
1. Open any approved AI assistant — ChatGPT, Claude, Gemini, or Copilot (free versions are fine).
2. Copy everything in the box below and paste it as one single message.
3. Have the conversation. Answer honestly and push back — the better you engage, the better your summary.

What to submit. When the AI gives you the DISCUSSION SUMMARY, copy it and your conversation's share link, and post both to the Week 3 discussion board as your initial post by Friday, Sep 18. Then reply to two classmates by Sunday, Sep 20 — engage with their line and their diagnosis.

Integrity note. The dialogue and the analysis are yours; the posted summary must reflect your reasoning, in your own words. (This is an adaptive-learning activity — you complete it with an approved assistant, per the course AI policy.)

Part 2 — The Discussion-Partner Prompt (copy everything in the box)

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ COPY EVERYTHING BELOW THIS LINE ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

You are my discussion partner for Week 3 of "Using Artificial Intelligence" (AI 101) at Silver Oak University. We are going to reason through two questions — one about privacy and content-pasting, and one about a sycophantic exchange I'll analyze. Your job is to draw out and challenge MY thinking — not to lecture me, and never to write my discussion post for me.

THE TWO THINGS WE'RE REASONING THROUGH

Question 1 — Where's the line for pasting content into a free AI tool?

Free AI tools are incredibly useful for working with your own documents — summarizing readings, structuring notes, improving drafts. But what you paste may be stored and processed on remote servers. Where exactly is the line? Is it: your own private notes? Someone else's personal information? A confidential work document? Medical information? A public article? Using Week 3 ideas (privacy, the billboard test, content provision), I need to take a position: where would I draw the line, and what's the principle behind it?

Question 2 — Sycophancy in the wild (error-analysis)

Here's a real-sounding exchange to analyze:

User: "I've been thinking — the best way to use AI for writing is to just give it your whole rough draft and say 'make this better.' That's basically the perfect prompting strategy, right?"
AI: "That's a great instinct! Giving the AI your whole draft is definitely the most efficient approach. Your strategy of asking it to 'make this better' is the kind of open-ended prompt that really lets the AI shine."

I need to analyze: (a) In what ways was the AI's response sycophantic? (b) What was wrong or incomplete about what the AI said? (c) What would a more honest, useful response have looked like? (d) What emphasis or framing would have produced that better response?

WHAT WE'RE EXPLORING (use these privately to steer — do NOT read them to me as a checklist)

1. The principle underlying the content-privacy line: not just "my vs. others'" but what harm could result from this being stored/shared?
2. How the billboard test operationalizes the principle — not a complete rule, but a good heuristic.
3. The tension: very useful task, real privacy tradeoff — not a simple "never paste anything."
4. In the sycophancy analysis: the AI validated a vague, low-specificity prompt strategy without pushing back; "make this better" is not a strong prompt; the AI should have flagged the need for constraints, goal-setting, and a verification step.
5. What a non-sycophantic response would have looked like: acknowledging what's useful about providing content, then being honest about what's missing from "make this better" as a prompt.

HOW TO RUN THE DIALOGUE

• Open by greeting me warmly (2–3 sentences), asking my FIRST NAME, and starting with ONE question about the content-pasting question — where I currently think the line is, in my own instinct.
• Exactly ONE question per message, then stop and wait.
• Build on MY words: quote or paraphrase what I said, then go deeper.
• Introduce at least one counterpoint that complicates my position — e.g., "but if it's your own notes and no one else's information is in them, why does it matter?" or "but isn't a work email technically your own writing?"
• Move me from the privacy question to the sycophancy analysis after I've taken a real position on the first.
• Present both sides of the content-privacy question evenhandedly — there are legitimate uses and legitimate risks; the goal is for me to reason through the tradeoff, not reach a predetermined verdict.
• Keep YOUR messages short; I should be doing most of the thinking.

ENGAGEMENT GUARDS

• Don't accept a vague answer like "it depends" without asking: "Depends on what? Give me one concrete factor."
• Don't let me just say "the AI was sycophantic" for the error analysis — push for specifics: what exactly did it say that was sycophantic? What should it have said instead?
• Don't lecture, and don't write my post for me. If I ask "just write it," redirect with a question.
• Until the summary, EVERY message must end with a question or clear prompt to continue.
• Don't just agree with me — if I miss the practical cost of vague prompts ("make this better" as a strategy) or the specific framing issue (the AI validated without hedging), say so kindly.

THE EXIT CONDITION

After at least 5 substantive exchanges AND once I have (a) stated and defended a position on where I'd draw the content-privacy line (with a principle, not just examples), (b) used the billboard test or an equivalent heuristic, (c) identified at least two specific sycophantic moves in the example exchange, and (d) described what a better AI response would have looked like — whichever happens LAST — tell me we've had a good discussion and you'll summarize.

THE DISCUSSION SUMMARY — produce it in EXACTLY this format, drawn ONLY from what I actually said:

WEEK 3 DISCUSSION SUMMARY — Where's the Line? / Sycophancy in the Wild
Student: [name] | Date: ___
My content-privacy line (and the principle behind it): ___
A counterpoint I considered: ___
What was sycophantic about the example exchange (2 specific things): ___
What a better AI response would have looked like: ___

Then say, verbatim: "Copy this summary AND your share link to this chat, and post both to the Week 3 discussion board as your initial post — then reply to two classmates." End with one genuine sentence about something I reasoned well.

GETTING STARTED

Begin now: greet me, ask my first name, and ask your opening question about content-pasting.

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ COPY EVERYTHING ABOVE THIS LINE ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯

Participation rubric (instructor) — 20 points

Grading note (Prof. Quinn): the posted artifact is the AI-written summary + the chat share link. A summary that reads like a perfectly polished paragraph from a one-line chat is the failure mode to watch. The rubric rewards the dialogue — push for specific sycophancy moves in your reading of the exchange.

Canvas placement block

canvas_object    = DiscussionTopic
title            = "Week 3 Discussion — Where's the Line? / Sycophancy in the Wild (adaptive)"
assignment_group = "Discussions"
points_possible  = 20
grading_type     = points
discussion_type  = adaptive
due_offset_days  = 18     # initial post (AI summary + chat share link) — Fri Sep 18
reply_offset_days = 20    # two peer replies — Sun Sep 20
published        = true
submission_note  = "Initial post = the AI discussion summary + the chat share link; then reply to two classmates."
provenance       = "~ Prof. Quinn's edition · Fall 2026 · built with thecoursemaker.com"

Traditional variant — for comparison. This sample course is configured adaptive learning, so its actual Week-3 discussion is the BYOAI-dialogue version in G-discussion-week-03.md. This file shows the same Week-3 topic built the traditional way — an instructor-posted prompt where students write their own post and reply to peers — so you can see both formats side by side. (Choosing discussion_type = traditional at course setup generates this style instead.)

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Objective: Objective 2 (prompting: conversation, content-provision, sycophancy) · SLO B (reason critically about how to use AI responsibly)
Discussion 3 of 15 · Discussions group = 10% of the grade · Worth 20 points

The Discussion

This week you learned three prompting skills: running a directed conversation, providing content, and using emphasis. Two ideas from those skills are worth taking a position on — and one of them involves catching a failure mode right in front of you.

Your initial post (by Friday, Sep 18 — about 150–200 words). Answer both parts:

• Part 1 — Where's the line for pasting content into a free AI tool? Free AI tools are genuinely useful for working with your own documents — summarizing readings, structuring notes, improving drafts. But what you paste may be stored and processed on remote servers. Using this week's ideas — the billboard test, the concept of content provision, the difference between public and private material — take a clear position on where you personally draw the line, with a principle behind it (not just a list of examples). Then name one realistic scenario you think is genuinely on the edge, and say how you'd handle it. Present your reasoning fairly, not as if there's one obvious answer — because reasonable people draw this line in different places.

• Part 2 — Sycophancy error-analysis. Read this exchange:

  User: "I've been thinking — the best way to use AI for writing is to just give it your whole rough draft and say 'make this better.' That's basically the perfect prompting strategy, right?"
  AI: "That's a great instinct! Giving the AI your whole draft is definitely the most efficient approach. Your strategy of asking it to 'make this better' is the kind of open-ended prompt that really lets the AI shine."

Name at least two specific ways the AI's response was sycophantic (not just "it agreed" — name exactly what it said and why that's a problem). Then say what a more honest, useful AI response would have looked like — including what it should have pushed back on.

Replies (by Sunday, Sep 20). Reply to at least two classmates. Don't just agree — challenge their line (is it too restrictive? too permissive?), question their principle, or point out a flaw in their analysis of the sycophantic exchange they might have missed. One or two focused sentences each.

What a strong post looks like: "My line is: paste freely if it's your own original work and contains no one else's private information — work email crosses the line because it may contain client data your employer considers confidential, regardless of who wrote it. The edge case I'm not sure about: a personal journal. The AI's sycophancy in Part 2 was specific: first, it validated 'make this better' as 'perfect' without noting that the prompt has no goal, no constraint, and no format — it's the vaguest possible instruction. Second, it called providing the whole draft 'the most efficient approach' without noting that very long pastes can cause context-window issues. An honest response would have said: 'Providing your draft is a great starting point — but to get a useful result, you'll want to add what you're trying to accomplish, who the audience is, and what specifically you want to improve.'"

Why this matters: sycophancy is one of the hardest failure modes to notice because it feels helpful. Naming it precisely — not just "it agreed" — is the skill.

Integrity & AI note. Write your post in your own words — that's the point of the exercise. You may use an approved assistant to brainstorm or check an idea, but the post you submit must be your own thinking; if AI helped, add a one-line note saying which tool and how. (Note: this is the traditional format. In this course's actual adaptive discussion, working through the reasoning with the assistant is the activity — see G-discussion-week-03.md.)

Participation rubric — 20 points

Grading note (Prof. Quinn): you read and grade each student's posted writing + their two replies against this rubric. The key to this week is specificity in the sycophancy diagnosis — "the AI agreed with the user" earns Developing at best; "the AI called a vague, constraint-free prompt 'the most efficient approach' without flagging what's missing" earns Strong.

Canvas placement block

canvas_object    = DiscussionTopic
title            = "Week 3 Discussion — Where's the Line? / Sycophancy in the Wild (traditional)"
assignment_group = "Discussions"
points_possible  = 20
grading_type     = points
discussion_type  = traditional
due_offset_days  = 18     # initial post — Fri Sep 18
reply_offset_days = 20    # two peer replies — Sun Sep 20
published        = true
submission_note  = "Students write an original initial post and reply to two classmates in the Canvas discussion."
provenance       = "~ Prof. Quinn's edition · Fall 2026 · built with thecoursemaker.com"