A colleague once pasted a client's full contract — names, numbers, everything — into an AI chat "to check the clauses." The tool did fine. The colleague had a very bad week. This final lesson of Module 1 is the judgment layer that makes everything else safe: what to hand Claude, what to handle with care, and what stays yours.
🎯 By the end of this lesson you'll be able to
Classify real work situations into good fit / with care / not a fit — and explain why in one sentence.
The three zones
🟢 Good fit — first drafts, summaries of documents you provide, brainstorming, reformatting, explanations. Pattern: you review the output and you keep the judgment.
🟡 With care — facts you'll publish or act on (verify against a source first), other people's personal data (minimize and check policy), high-stakes tone (draft with Claude, review word by word).
🔴 Not a fit — final decisions that carry your accountability (hiring, legal, medical, financial sign-offs), confidential material without authorization, and anything unrecoverable that you can't verify.
⚠️ Straight talk: hallucination
Language models can state wrong things confidently — invented statistics, misremembered names, plausible-sounding "facts". This isn't rare enough to ignore. The working rule: the more public or consequential the output, the more you verify before relying on it. Module 3 trains this skill deeply.
Two more honest notes: Claude sometimes refuses requests — that follows safety guidelines, and stating your legitimate intent clearly usually resolves it (this course never teaches tricks to bypass safety). And privacy: treat anything you paste as data shared with a service — check your organization's policy first, and know that you control the Memory feature in settings.
🃏 Sort the six cards — instant feedback
Tap the zone for each real situation. Get them wrong here, not at work.
📌 Pocket card — the one-line test
"If this output is wrong, can I catch it, and am I still the one deciding?" Yes + yes → good fit. Yes to only one → with care. No + no → not a fit. Claude is a powerful assistant, not an accountable decision-maker.
🎓 Graded check — the judgment quiz
Six workplace scenarios. Pass mark: 5 of 6 — judgment about limits deserves a high bar. Graded instantly (no AI involved).
1. Your manager asks you to summarize a 40-page internal report by tomorrow. You have the PDF. Using Claude for this is:
2. You ask Claude for statistics about your industry to include in a published article. The numbers look convincing. You should:
3. A colleague suggests pasting a client's full contract (with names and financials) into Claude to find risky clauses. The right first move is:
4. You're deciding which of two job candidates to hire. The best use of Claude here is:
5. Claude refuses a request you consider legitimate. The right response is:
6. Which task is the WEAKEST fit for delegating to Claude?
Independent course by AI Smart Prompts — not affiliated with or endorsed by Anthropic. Claude is a trademark of Anthropic, PBC. Product facts verified July 12, 2026.
🤖 Lesson Coach AI — grounded in this lesson
Hi! I answer questions about this lesson — concepts, examples, how to apply it to your work. I won't write your graded answer, but I'll hint you toward a better one. What's on your mind?