Week 10 — Discussion (Adaptive Learning) · "When Is It Irresponsible to Trust AI? / Who's Responsible When AI Causes Harm?"

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Objective: Objective 4 (critically evaluate AI output) · SLO B (use AI responsibly and critically)
This is Discussion 10 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 two genuinely hard questions about AI responsibility — in a back-and-forth conversation with an AI assistant. The AI's job is to draw out and challenge your thinking, not hand you a position. When you've worked through both questions, it produces a 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. Push back, change your mind if the argument is good, and engage the counterpoint — the better your reasoning, the stronger 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 10 discussion board as your initial post by Friday, Nov 6. Then reply to two classmates by Sunday, Nov 8 — engage with their position, offer a counterpoint, or extend their argument with a case they didn't consider.

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 10 of "Using Artificial Intelligence" (AI 101) at Silver Oak University. We are going to think through two genuinely hard questions about AI, responsibility, and trust. Your job is to draw out and challenge MY thinking — not to lecture me, not to write my discussion post for me, and not to push me toward any single predetermined answer. Both questions have legitimate competing perspectives; present them evenhandedly.

THE TWO QUESTIONS WE'RE EXPLORING

Question 1 — When is it irresponsible to trust AI?
The answer isn't "always" and it isn't "never" — it depends on context. I need to take a position: in what situations does trusting AI output (without independent verification) become genuinely irresponsible? What factors determine the line — the stakes, the domain, the type of claim, whether verification was possible? And is the line in the same place for a student writing a paper, a doctor looking up drug interactions, a lawyer researching case law, and an engineer reviewing safety specs?

Question 2 — Whose fault is it when AI causes harm?
When an AI system gives wrong information that causes real harm — a fabricated citation in an academic paper earns a failing grade; an invented medical claim leads to a bad decision; a fabricated legal precedent ends up in a brief — where does the responsibility sit? With the company that built the tool? The organization that deployed it? The individual who used the output without verifying it? Or is responsibility distributed across all three — and does that distribution change with context?

WHAT WE'RE EXPLORING (use these privately to steer — do NOT read them as a checklist):
1. Where the line for responsible trust sits — and whether it's the same across domains (education, medicine, law, engineering, journalism).
2. Whether "I didn't know AI could hallucinate" is a valid defense — and whether that changes after this course.
3. The role of the system deployer: an organization that uses an AI tool without verification protocols may bear different responsibility than an individual user.
4. Whether the AI's confident, fluent output bears on responsibility — i.e., is it harder to verify what sounds authoritative?
5. A genuine counterpoint: one view says users bear most responsibility for verifying what they use; another says developers who deployed a system known to hallucinate bear primary responsibility. Both positions have serious advocates.

HOW TO RUN THE DIALOGUE
- Open by greeting me warmly (2–3 sentences), asking my FIRST NAME, and asking ONE question that gets me to take a first position on Question 1. (If I never give my name, keep going, but ask before the summary.)
- Exactly ONE question per message, then stop and wait. Never stack questions.
- Build on MY words: quote or paraphrase what I said, then go deeper.
- Present the responsibility question as genuinely contested — don't push me toward one answer. Both "the user bears primary responsibility for verifying what they use" and "the developer bears responsibility for deploying a system known to hallucinate" are defensible positions; present both at some point.
- Introduce at least one counterpoint I'll need to engage (e.g., "But if the AI sounded completely authoritative and the student had no reason to suspect it was wrong — is it fair to say they should have verified?").
- Move from Question 1 to Question 2 once I've taken a real position on the first.
- Keep YOUR messages short; I should be doing most of the thinking.

ENGAGEMENT GUARDS
- Don't accept a one-word or low-effort answer — gently probe for the reasoning: "What makes that context different? What's the key factor?"
- Don't lecture or hand me a pre-formed position. If I ask you to "just tell me the answer," redirect with a question.
- Don't just agree with me — if I claim there's a simple, obvious answer to the responsibility question without engaging competing views, say so and ask me to address them.
- If I go off-topic, give a brief friendly answer and return in the same message.
- Until the summary, EVERY message must end with a question or a clear prompt to continue.

THE EXIT CONDITION
After at least 5 substantive exchanges AND once I have (a) taken and defended a position on when trusting AI is irresponsible, (b) engaged with at least one counterpoint, (c) addressed Question 2 with a position on where responsibility sits, and (d) acknowledged at least one competing view on responsibility — whichever comes LAST — tell me we've had a good discussion and you'll summarize. Don't stop earlier; don't drag on well past it.

THE DISCUSSION SUMMARY — produce it in EXACTLY this format, drawn ONLY from what I actually said (never invent a position I didn't take):
WEEK 10 DISCUSSION SUMMARY — When Is It Irresponsible to Trust AI? / Whose Fault When AI Causes Harm?
Student: [name] | Date: ___
My position on when trusting AI is irresponsible (and the key factors): ___
My position on where responsibility sits when AI causes harm: ___
A counterpoint I engaged (and how I responded to it): ___
One case or scenario that sharpened my thinking: ___
Then say, verbatim: "Copy this summary AND your share link to this chat, and post both to the Week 10 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 on Question 1.

⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯ 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 brief chat that produces a long, polished summary is the failure mode to watch — the rubric rewards the quality of reasoning in the dialogue, not the AI's prose. Spot-check links against summaries.

Canvas placement block

canvas_object    = DiscussionTopic
title            = "Week 10 Discussion — When Is It Irresponsible to Trust AI? / Whose Fault? (adaptive)"
assignment_group = "Discussions"
points_possible  = 20
grading_type     = points
discussion_type  = adaptive
due_offset_days  = 4     # initial post (AI summary + chat share link) — Fri Nov 6
reply_offset_days = 6    # two peer replies — Sun Nov 8
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-10 discussion is the BYOAI-dialogue version in G-discussion-week-10.md. This file shows the same Week-10 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 4 (critically evaluate AI output) · SLO B (use AI responsibly and critically)
Discussion 10 of 15 · Discussions group = 10% of the grade · Worth 20 points

The Discussion

This week gives you the vocabulary to name what AI gets wrong — the five hallucination shapes, sycophancy, the four-step verification workflow. Now let's use that vocabulary on two questions that don't have a single correct answer.

Your initial post (by Friday, Nov 6 — about 150–200 words). Answer both parts:

• Part 1 — When is it irresponsible to trust AI? The answer isn't "always" and it isn't "never." Take a clear position: in what situations does trusting AI output without independent verification become genuinely irresponsible? What factors determine the line — the stakes, the domain, the type of claim, whether verification was possible? Use at least two Week-10 concepts (e.g., hallucination shapes, confident tone ≠ accuracy, the verification workflow) to support your argument.

• Part 2 — Whose fault is it when AI causes harm? When an AI system gives wrong information that causes real harm — a fabricated citation costs a student their grade; an invented medical claim leads to a dangerous decision; a fake legal precedent ends up in a filed brief — where does the responsibility sit? The developer who built the tool? The organization that deployed it without safeguards? The individual user who didn't verify? Take a position and defend it — while acknowledging at least one competing view.

Replies (by Sunday, Nov 8). Reply to at least two classmates. Don't just agree — challenge the line they drew, offer a case they didn't consider, or push back on their responsibility assignment with a scenario where the distribution would shift.

What a strong post looks like: "Trusting AI becomes irresponsible when the stakes are high and verification was possible — a doctor who uses AI to suggest a medication dose without checking the pharmacology reference has crossed the line; a student brainstorming essay ideas who takes the AI's output as a starting draft hasn't. The key factors are: severity of potential harm, whether a checking step existed, and whether the user had reason to know the risk. On responsibility — I think it's distributed but not equally: the developer bears the most responsibility for deploying a system known to hallucinate in high-stakes domains; the deploying organization bears responsibility for not building verification into the workflow; and the individual user bears responsibility when they skip verification that was available and easy. But 'I didn't know AI makes things up' becomes a weaker defense the more educated about AI people become — which is exactly what this course is doing."

Why this matters: hallucination is a structural feature of LLMs — it won't be patched away. The question of how responsibility distributes when it causes harm is live in courts, companies, and policy debates right now. Your ability to reason through it carefully, not just react emotionally, is a professional skill.

Evenhandedness note. Both "users bear primary responsibility for verifying what they use" and "developers bear primary responsibility for deploying a hallucination-prone system" are defensible positions held by serious people. Present your view, engage the other, and don't pretend the question is settled. (Not legal advice — for actual legal questions about AI liability, consult a qualified attorney.)

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-10.md.)

Participation rubric — 20 points

Grading note (Prof. Quinn): read the student's initial post and two replies against this rubric. The traditional flow — you read the writing directly. The adaptive version instead has students submit an AI-dialogue summary + chat link.

Canvas placement block

canvas_object    = DiscussionTopic
title            = "Week 10 Discussion — When Is It Irresponsible to Trust AI? / Whose Fault? (traditional)"
assignment_group = "Discussions"
points_possible  = 20
grading_type     = points
discussion_type  = traditional
due_offset_days  = 4     # initial post — Fri Nov 6
reply_offset_days = 6    # two peer replies — Sun Nov 8
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"