Chapter 32 Exercises: When NOT to Use AI (and Why That Matters)
Instructions
These exercises combine reflective analysis with practical workflow design. They are designed to move you from theoretical understanding of AI no-fly categories to a concrete, personalized no-fly list that you can reference in your actual work. Complete them in order — they build toward the final protocol exercise.
Part 1: Category Analysis
Exercise 1.1: Safety-Critical Assessment
For your specific professional domain, identify three tasks that involve or border on safety-critical outcomes. For each:
a) What is the potential harm if the task produces an incorrect result? b) Is AI currently used (by you or in your field) for this task? c) What human oversight or verification safeguard currently exists? d) Is that safeguard adequate given AI's known failure modes (hallucination, confident error)?
Be honest about your current practice, not just your aspirational practice.
Exercise 1.2: The Authenticity Test
For each type of communication below, evaluate: Would this communication be diminished if the recipient knew AI wrote it? Rate each: Definitely yes / Probably yes / Depends on context / Probably not / Definitely not.
- A condolence message to a close colleague who lost a parent
- A "thank you for your interest but you didn't get the job" email to a candidate
- A reference letter for a student you mentored
- A cold outreach email to a potential new client
- A birthday message to a close friend
- An apology to a client for a project delay
- A LinkedIn recommendation for a former colleague
- An automated out-of-office email response
What patterns emerge from your ratings? What principle explains where you drew the lines?
Exercise 1.3: Learning Context Audit
Make a list of the professional skills that are important for your career over the next three to five years. For each:
a) Am I currently practicing this skill in AI-free contexts? b) Have I noticed any degradation in this skill since I started using AI assistance? c) Is there an AI-free practice zone I should create for this skill?
Be honest about skills you may have been delegating to AI that you need to keep sharp.
Part 2: Confidentiality Assessment
Exercise 2.1: Information Audit
List the five most sensitive types of information you regularly work with in your professional role. For each:
a) What confidentiality obligations apply? (legal privilege, HIPAA, NDA, trade secret, other) b) Have you or your team members ever input this type of information into a consumer AI tool? c) What is your current policy, explicit or de facto? d) Is your current policy adequate given the confidentiality obligations?
Exercise 2.2: The Public Disclosure Test
For each of the following, apply the test: "Would I be comfortable if this information became publicly visible?"
a) A client's marketing strategy you've been helping develop b) A patient case description you're asking AI to help summarize c) Internal company performance data you're asking AI to help analyze d) A draft of a contract you're asking AI to help review e) An HR situation involving a specific named employee you're asking AI for advice on
For any case where the answer is "no," what does that tell you about appropriate tool selection?
Exercise 2.3: Enterprise vs. Consumer Tool Decision
Research the data handling policies of the AI tools you currently use. For your primary AI tool:
a) What does the tool's privacy policy say about data retention and training use? b) Is there an enterprise tier with different data handling commitments? c) Does your organization have a policy on which types of information can go into which AI tools? d) If your organization doesn't have such a policy, should you draft one?
Part 3: Skill Atrophy Analysis
Exercise 3.1: The Before/After Self-Assessment
Think about your professional capabilities one year ago vs. today. For skills where you have been using AI assistance:
a) Are those skills stronger, the same, or weaker than they were a year ago? b) In skills that have weakened: is AI use a contributing factor? c) In skills that have strengthened: is AI use a contributing factor?
Be genuinely honest — this assessment is for your own benefit.
Exercise 3.2: The "Without AI" Stress Test
Choose three tasks that you now routinely use AI for. Perform each one without AI assistance. Observe:
a) How much harder was it without AI? b) Was the difficulty due to (i) appropriate task complexity, (ii) skill degradation from delegation, or (iii) unfamiliarity with doing it without AI even if the skill is intact? c) What do your observations tell you about skill maintenance in these areas?
Exercise 3.3: Designing AI-Free Practice Zones
Based on your analysis in Exercises 3.1 and 3.2, identify two to three skills that need AI-free practice zones — specific, regular tasks you will perform without AI to maintain or develop those skills.
For each: - What specific task type will you do AI-free? - How often? - How will you know if the practice is working?
Part 4: The "Just Because You Can" Problem
Exercise 4.1: Motive Examination
For the AI uses you have recently applied in professional contexts, examine your motive for each:
a) Was AI the right tool for this task? Or was it just the easiest available option? b) Did using AI serve the actual purpose of the task, or just the surface-level output requirement? c) Is there a pattern in which tasks you use AI for where AI is not actually the right fit?
Exercise 4.2: The Recipient's Perspective
For three pieces of AI-assisted content you have produced recently, consider:
a) What would the recipient of this content think if they knew AI contributed to it? b) Would that change their relationship to the content? To you? c) Would the disclosure change anything about whether the use was appropriate?
Exercise 4.3: The Relationship Capital Question
Professional reputation is built through consistent demonstration of genuine capability, judgment, and care. Identify two ways in which your AI use might be — even subtly — trading relationship capital for short-term efficiency. What would you do differently?
Part 5: Building Your Personal AI No-Fly List
Exercise 5.1: The Core List
Using the six questions from Section 8 as your framework, work through your professional and personal life systematically. For each major domain (professional tasks, client/colleague communication, personal relationships, learning and development, confidentiality-bearing information):
For each domain: 1. List the task types in that domain 2. Apply the six questions to each 3. Mark tasks where any answer is "yes" as candidates for the no-fly list
After completing all domains, review your candidates and draft your preliminary no-fly list.
Exercise 5.2: The No-Fly List Format
A useful personal no-fly list includes: - The specific task or context (not just a category, but a concrete description) - The reason it's on the list (which of the six questions it triggered, and why) - The alternative approach (if not AI, then what?) - The review trigger (what would cause you to reconsider this boundary?)
Format your list with these four elements for each item.
Exercise 5.3: The List Review with a Trusted Colleague
Share your no-fly list (or a subset of it) with a trusted colleague or mentor who also uses AI tools professionally. Ask them:
a) Does anything on the list surprise you? b) Is there anything you think I'm being too restrictive about? c) Is there anything you think I'm not restrictive enough about? d) Is there anything I missed that should be on the list?
The goal is not to get external validation of every decision, but to test your reasoning against someone else's perspective.
Part 6: The Nuanced Middle Ground
Exercise 6.1: The AI Assistance Spectrum
For each of the following, place the activity on a spectrum from "appropriate AI assistance" to "inappropriate AI replacement":
a) Asking AI to fix grammar and clarity in a cover letter you wrote b) Asking AI to write a cover letter that you lightly edit c) Asking AI to write a cover letter from scratch that you submit unchanged d) Asking AI to help you identify what to say in a personal statement e) Asking AI to write your personal statement, which you review and approve f) Submitting an AI-generated personal statement without reading it carefully
Where are the meaningful distinctions? What principle separates acceptable from unacceptable in this spectrum?
Exercise 6.2: The Disclosure Variable
For five situations on your no-fly list or close to it, ask: If I disclosed AI involvement to the relevant party, would the use become acceptable?
What does this exercise reveal about the role of transparency in determining appropriateness?
Part 7: Reflection
Exercise 7.1: Your Relationship With AI
Reflect on your current AI use. Write a one-page honest assessment of: - Where AI adds genuine value in your professional life - Where you may have extended AI use beyond where it adds value - What you are prepared to do differently based on what you've learned in this chapter
Exercise 7.2: The No-Fly List as Identity Statement
Your no-fly list implicitly states something about your professional values: what you believe authentic work looks like, what skills you want to maintain, what obligations you take seriously, what relationships you won't transact through a model.
Write a brief paragraph that articulates what your no-fly list says about your professional identity and values.