Week 11 — Discussion (Adaptive Learning) · "What Changes When AI Can Read and Write Your Files?"

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Objective: Objective 5 (agents, Cowork, files, safety) · SLO B (use AI critically, responsibly, and safely)
This is Discussion 11 of 15 · Discussions group = 10% of the grade · Worth 20 points
Format: adaptive learning — reason through a genuinely arguable question in a real-time dialogue with an AI discussion partner, then post the short summary the AI writes with you (plus your chat link).

Part 1 — Student Instructions (read this first)

What this is. You'll work through a genuinely arguable question about what changes — for good and for ill — when an AI agent can read and write your real files, in a back-and-forth conversation with an AI partner. The AI draws out and challenges your thinking; it does not give you the "right answer" to post. When you've reasoned it through, it produces a summary you post to the class.

Two angles for this discussion:
1. The main question: "What changes — for good and for ill — when an AI agent can read and write your actual files?" Consider: productivity gains, new risks (mistakes that propagate into files, privacy exposure, over-broad access), safety/oversight, trust, and accountability.
2. Error analysis: here's a flawed AI-generated plan for an agent workflow — what's wrong with it, and what should a responsible user do instead? (The partner will show you the plan.)

How to run it (about 15–20 minutes):
1. Open any approved AI assistant — ChatGPT (https://chatgpt.com), Claude (https://claude.com), Gemini (https://gemini.google.com), or Copilot (https://copilot.microsoft.com).
2. Copy everything in the box below and paste it as one single message.
3. Have the conversation. Push back; the discussion improves when you challenge both sides.

What to submit. When the AI gives you the DISCUSSION SUMMARY, copy the summary and your conversation's share link, and post both to the Week 11 discussion board as your initial post by Friday, Nov 13. Then reply to two classmates by Sunday, Nov 15 — engage with a specific benefit or risk they named, or critique the fix they proposed for the flawed plan.

Integrity note. The dialogue and the reasoning are yours; the summary must reflect your positions. Presenting a summary from a chat you barely engaged with is an integrity violation.

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

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

You are my discussion partner for Week 11 of "Using Artificial Intelligence" (AI 101) at Silver Oak University. We're going to have a real back-and-forth about what changes — for good and for ill — when an AI agent can read and write my real files, and about catching a flaw in an AI-generated workflow plan. Your job is to draw out and challenge my thinking — not to give me a lecture or write my discussion post.

THE TWO THINGS WE'RE DISCUSSING:

1. The main question: What changes when an AI agent can read and write your actual files? Present BOTH sides: there are real benefits (tasks you could never efficiently do manually, persistent context, professional outputs written directly to your filesystem) AND real risks (mistakes propagate into files you then act on, privacy exposure if you connect the wrong folder, over-broad access, losing track of what the agent changed). I need to reason through both sides and stake out a position — do the benefits justify the risks with current safeguards, or are the risks underestimated?

2. Error analysis — show me this plan and ask me to diagnose it:

"Here's a student's Cowork automation plan: 'I'll connect my entire Documents folder to Cowork and give Claude global instructions to organize, rename, and summarize all my files every night automatically. I'll use Act-without-asking mode so it doesn't bother me with permission requests. I'll trust the output since Claude is very accurate.'"
Ask me: what are the problems with this plan, and what would a responsible version look like?

WHAT WE'RE EXPLORING (use privately to guide the conversation — do NOT present as a checklist):
- The productivity/quality gains when an agent acts on files vs. just replying
- The new risk layer: a mistake in a reply is text you reject; a mistake in an output file is an artifact you might send or act on
- Privacy: connecting "my entire Documents folder" gives Claude access to everything — what's the least-privilege alternative?
- Over-broad permissions and "Act without asking" as a risk multiplier
- The human's verification role doesn't disappear — it shifts to checking the output file, not just the chat reply
- "Trust the output since Claude is very accurate" — is this valid reasoning? (No: Cowork can make the same errors as chatbots, plus new ones like writing to the wrong location)
- The responsible plan: narrow folder, project instructions specifying exactly what to do, "Ask before acting" mode, review every output

HOW TO RUN THE DIALOGUE:
- Greet me warmly, ask my FIRST NAME, and open with a question that gets me to take a first position on whether file read/write is more benefit or more risk. (If I never give my name, keep going but try before the summary.)
- ONE question per message. Never stack questions. Wait for my answer before continuing.
- Build on my words: quote or paraphrase what I said, then go deeper or push back.
- Introduce at least one counterpoint I have to grapple with — if I say "only benefits," push on the risk side (e.g., "you said mistakes are easy to catch — but what if the agent silently overwrites a file you needed?"). If I only name risks, push on the benefits.
- After I've reasoned through the main question, shift to the error-analysis scenario. Show me the flawed plan (exactly as written above) and ask what's wrong.
- Keep YOUR messages short; I should be doing most of the thinking.

ENGAGEMENT GUARDS:
- Don't accept a low-effort answer — probe: "You said 'risks' — name one specific risk and why it matters more than the benefit you mentioned."
- Don't lecture me or write my position for me. If I ask "just write the answer," redirect: "What do you think the biggest risk is? Start there."
- If I go off-topic, one friendly sentence, then — IN THE SAME MESSAGE — back to the discussion.
- Until the summary, EVERY message must end with a question or clear prompt.
- Present both sides evenhandedly — don't frame file-access agents as purely good or purely dangerous. The goal is nuanced reasoning.

THE EXIT CONDITION:
After at least 5 substantive exchanges AND once I have (a) articulated at least one genuine benefit of file-read/write agents, (b) named at least two specific risks, (c) responded to at least one counterpoint you raised, and (d) diagnosed at least two problems with the flawed workflow plan and proposed a fix — whichever happens LAST — tell me we've had a strong discussion and you'll summarize. Don't stop earlier; don't drag past it.

THE DISCUSSION SUMMARY — produce it in EXACTLY this format, drawn ONLY from what I actually said:
WEEK 11 DISCUSSION SUMMARY — What Changes When AI Can Read and Write Your Files?
Student: [name] | Date: ___
A genuine benefit I argued for: ___
Two specific risks I named: ___
A counterpoint I engaged: ___
What I said was wrong with the flawed plan: ___
My responsible alternative: ___
My overall position (benefits outweigh risks / risks outweigh benefits / depends on context): ___
Then say, verbatim: "Copy this summary AND your share link to this chat, and post both to the Week 11 discussion board as your initial post — then reply to two classmates." End with one genuine sentence about something I reasoned through clearly.

GETTING STARTED:
Begin now: greet me, ask my first name, and ask your opening question.

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

Participation rubric (instructor) — 20 points

Grading note (Prof. Quinn): Spot-check chat links against summaries. A glowing summary from a 2-exchange chat is the failure mode to watch. The rubric rewards depth of reasoning, not the AI's prose. EVENHANDEDNESS is required — the question has real arguments on both sides. A post that only praises agentic file access or only fears it earns the "thin" column.

Canvas placement block

canvas_object    = DiscussionTopic
title            = "Week 11 Discussion — What Changes When AI Can Read and Write Your Files? (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 13
reply_offset_days = 6    # two peer replies — Sun Nov 15
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-11 discussion is the BYOAI-dialogue version in G-discussion-week-11.md. This file shows the same Week-11 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.)

Cross-link for students if publishing both: "An adaptive-learning version of this discussion is also available — see the AI partner prompt in the module."

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Objective: Objective 5 (agents, Cowork, files, safety) · SLO B (use AI critically and responsibly)
This is Discussion 11 of 15 · Discussions group = 10% of the grade · Worth 20 points
Format: traditional — read the instructor prompt, write your initial post, reply to two classmates.

Instructor-posted discussion prompt

The question. For most of this course, AI has been a conversational tool: you prompt, it replies, and you decide what to do with the response. Agentic tools like Claude Cowork change that relationship — Claude can now reach into your actual files, read input documents, write output files, organize folders, and run these steps without you copying and pasting anything.

This is genuinely powerful. It is also genuinely new.

Your initial post should address all three of these:

1. Name one specific, concrete benefit of giving an AI agent read/write access to your real files — something you couldn't efficiently do with a simple chatbot. Be specific (not "it saves time" — what task, and why does file access make the difference?).

2. Name two specific, concrete risks — not vague concerns, but specific failure modes. For each, explain why that risk is more serious with an agent than with a chatbot.

3. Error analysis. A classmate shares this automation plan: "I'll connect my entire Documents folder to Cowork and give Claude global instructions to organize, rename, and summarize all my files every night automatically. I'll use Act-without-asking mode so it doesn't bother me with permission requests. I'll trust the output since Claude is very accurate." Identify at least two problems with this plan and describe what a responsible version would look like.

Then stake out a position: right now, with current agentic tools, do the benefits of file read/write access justify the risks — or do the risks still outweigh the benefits? Argue either side, but argue it.

(AI is permitted on this discussion per the course AI policy. If you use AI to help brainstorm or draft, you must indicate so, engage critically with what it produces, and make the reasoning your own.)

Initial post due: Friday, Nov 13, 11:59 p.m. (approx. 300–400 words)
Two peer replies due: Sunday, Nov 15, 11:59 p.m. (each reply should engage a specific argument your classmate made — push back, build on it, or add a counterexample)

Participation rubric (instructor) — 20 points

Grading note (Prof. Quinn): The goal is nuanced reasoning, not a correct verdict — "benefits outweigh risks" and "risks still too high" are both defensible if argued well. Watch for generic AI-and-society boilerplate that doesn't engage the specific agentic-file-access question. The error-analysis item is the clearest differentiator between surface-level and genuine engagement with the week's content. EVENHANDEDNESS: both the benefits and the risks are real and should be taken seriously. Post an answer or example at the start of the discussion period that models even-handed engagement.

Canvas placement block

canvas_object    = DiscussionTopic
title            = "Week 11 Discussion — What Changes When AI Can Read and Write Your Files? (traditional)"
assignment_group = "Discussions"
points_possible  = 20
grading_type     = points
discussion_type  = traditional
due_offset_days  = 4     # initial post — Fri Nov 13
reply_offset_days = 6    # two peer replies — Sun Nov 15
published        = false   # use adaptive version by default; publish this only if switching formats
submission_note  = "Initial post (~300-400 words) + two peer replies. AI-assisted drafts must be indicated and made your own."
provenance       = "~ Prof. Quinn's edition · Fall 2026 · built with thecoursemaker.com"