🤖 Part 24: Ethics, Integrity & When Not to Use AI
Powerful tools come with responsibilities, and AI assistants are no exception. The good news is that using AI ethically isn't complicated or gloomy — it's mostly common sense plus a few habits worth naming out loud. This part walks through integrity, bias, over-reliance, misinformation, and privacy in a balanced, practical way, and ends with the most useful list of all: the moments when the right choice is simply not to use AI.
📚 What You'll Learn
By the end of this part, you'll understand:
- How to handle academic and workplace integrity — disclosure, following the rules, and not passing AI work off as your own where that's prohibited
- The real risks to watch for: bias, over-reliance, misinformation and deepfakes, and other people's privacy
- A brief, balanced note on the broader context — labor and the environment — without the doom or the hype
- A clear, practical "when NOT to use AI" list you can actually apply
In This Part
Integrity: Whose Work Is It?
The most common ethical question people meet is also the most practical: when is it fine to use AI for something, and when do you need to say so — or hold back entirely? There's no single universal rule, because the answer depends on the setting and its expectations. But a few principles travel well and will keep you out of trouble almost everywhere.
📖 Definition
Integrity, in AI terms: being honest about how your work was made. It doesn't mean never using AI — it means not misrepresenting AI-assisted work as entirely your own where that representation matters, such as in schoolwork or a job where the rules require otherwise.
Start with the rules that apply to you. Schools and employers are writing AI policies quickly, and they vary a lot: some encourage AI with disclosure, some restrict it to certain tasks, some ban it for graded or official work. Your first move in any serious setting is to find out what's actually allowed — ask a teacher, read the syllabus, check the employee handbook — rather than guessing.
| Situation | The honest move |
|---|---|
| A graded assignment or exam | Follow your school's policy exactly; if AI is limited or banned, don't use it — and if unsure, ask before, not after |
| Work your employer will rely on or publish | Check the workplace policy; disclose your AI use if that's expected, and never let it create a false impression of who did what |
| Your own personal projects | Use AI freely — it's your call — while still verifying anything you'll act on |
| Something with your name on it as author | You're accountable for every word, AI-assisted or not; own it, check it, and be ready to stand behind it |
⚠️ Important: "The AI wrote it" is never a defense. Whether it's a fabricated citation in an essay, a wrong figure in a report, or an offhand claim in an email, you put your name on it, so you are responsible for it. Using AI raises the bar on checking your work, it doesn't lower the bar on owning it.
✅ Tip — disclosure is usually easy and disarming
When in doubt, a simple line does the job: "I used an AI assistant to draft this and then edited and verified it." Most people appreciate the honesty, and it costs you nothing. If a setting would frown on that disclosure, that's a strong signal AI wasn't appropriate there in the first place.
Bias, Misinformation & Deepfakes
Because an assistant learned from an enormous amount of human writing, it also absorbed humanity's biases — the stereotypes, blind spots, and skews baked into that text. It doesn't hold prejudices the way a person does, but it can reproduce them: default to certain assumptions about who does what job, lean toward one culture's perspective, or handle some groups less fairly than others. Companies work to reduce this, but no assistant is free of it.
⚠️ Watch Out — bias hides in "neutral" output
Bias rarely shows up as anything obvious. It's in the example names an assistant picks, the assumptions it makes when details are missing, whose viewpoint it centers, and what it treats as the "normal" case. When you're writing about or for people, read with that in mind — and when representation matters, ask directly for a broader range of perspectives or examples.
The second risk you already know from earlier parts, now with a sharper edge: assistants can produce misinformation — confident, fluent, and wrong — and they can help generate deepfakes, convincing fake text, images, audio, or video. That raises two duties. First, keep verifying what you receive so you don't pass along something false. Second, don't create deceptive content: fake reviews, impersonations, fabricated quotes, or images meant to trick people are harmful whether a human or an AI made them.
or harm someone?"} B -->|"No — low stakes"| C["✅ Use it, lightly checked"] B -->|"Maybe — facts matter"| D["🔎 Verify before sharing"] B -->|"Yes — deceptive by design"| E["🚫 Don't create or spread it"]
🧠 Mindset
A useful test before you publish or forward anything AI helped make: "Would I be comfortable if the person affected knew exactly how this was made?" If yes, you're almost certainly fine. If the honest answer makes you wince, that's your signal to slow down, verify, disclose, or reconsider.
Over-Reliance, Privacy & the Bigger Picture
Not every risk is dramatic. One of the quieter ones is over-reliance: if you hand every piece of thinking to an assistant, the underlying skills — writing clearly, reasoning through a problem, doing the arithmetic yourself — can slowly fade from disuse. AI is a fantastic collaborator and a poor replacement for your own judgment. Use it to go faster and further, not to stop thinking. A good check: could you still do this without the assistant if you had to?
Then there's other people's privacy. Earlier parts covered protecting your own sensitive information; the flip side is respecting everyone else's. Be careful about pasting other people's personal details, private messages, medical or financial information, or confidential work into an assistant. The fact that something is convenient to feed in doesn't mean it's yours to share.
⚠️ Watch Out — a good rule for other people's data
Before you paste something that involves another person, ask whether they'd be comfortable with it going into an AI tool. Names, contact details, private conversations, health or money matters, and anything shared with you in confidence deserve real caution — strip identifying details, or don't paste them at all.
Finally, a brief and balanced word on the bigger picture, because a fair guide shouldn't skip it or preach about it. There are real, unsettled conversations around AI's effect on jobs and labor — which work changes, which disappears, which is created — and around its environmental footprint, since training and running large models uses meaningful energy and water. Thoughtful people disagree on the scale and the trade-offs. You don't have to resolve those debates to use these tools; it's simply worth being an aware, intentional user rather than a thoughtless one.
✅ Tip — awareness, not guilt
The goal here isn't to make you feel bad for using AI — these are genuinely useful tools, and using them is reasonable. It's to use them on purpose: reach for AI when it truly helps, skip it when it doesn't add much, and stay curious about the wider effects. Intentional beats both anxious and oblivious.
When NOT to Use AI
The most practical ethics lesson isn't a philosophy — it's a short list of situations where the responsible move is to not rely on an AI assistant, or to insist a qualified human stays firmly in the loop. Keep this one handy.
| Don't lean on AI for… | Why, and what to do instead |
|---|---|
| High-stakes decisions with no human in the loop | A confident wrong answer can do real damage; use AI to inform, but let a responsible human make and own the call |
| Serious medical questions | It can explain terms and help you prepare questions, but it's not a doctor; for diagnosis or treatment, see a professional |
| Legal matters that matter | Fine for plain-language background; for actual advice, contracts, or filings, consult a qualified lawyer |
| Financial decisions of consequence | Helpful for learning concepts; for real money moves, talk to a licensed professional |
| Anything you can't verify but must rely on | If you can't check it and the stakes are real, don't treat AI output as fact — find an authoritative source |
| Crisis or safety situations | In an emergency, contact real emergency services or a qualified human — not a chatbot |
⚠️ Important: The common thread is simple — the higher the stakes, the more a qualified human matters. AI can be a wonderful assistant to a doctor, a lawyer, an advisor, or a careful decision-maker. It's a dangerous replacement for one. Let it help; don't let it decide the things that truly count.
🧠 Mindset
None of this makes AI off-limits for serious topics — you can absolutely use it to understand your diagnosis, prepare for a legal meeting, or learn how a mortgage works. The line is between informing yourself (great) and substituting for a professional on a decision that carries real consequences (not great). Keep that line clear and you'll navigate almost every hard case correctly.
🛠️ How To: Write Your Own Responsible-Use Card
What you'll do: write a short, personal "responsible AI use" card — a handful of lines you can keep near your desk or in a note, so good habits become automatic rather than something you reason out each time.
Step by step
- Write your verify rule. One line, in your own words — for example, "I verify anything I'll act on, publish, or send to someone else."
- Write your disclosure rule. Decide when you'll say you used AI — for schoolwork, client work, or anything with your name on it — and phrase it as a habit you'll follow.
- Write your privacy rule. Note what you won't paste: your own sensitive data, and other people's personal, confidential, or private information.
- List your personal "not for AI" lines. Copy the situations from this part that apply to your life — medical, legal, financial, high-stakes, unverifiable — in your own words.
- Add one integrity check. A single question to ask yourself, such as "Would I be fine if the person affected knew exactly how this was made?"
- Keep it visible. Pin it in a note app or on paper where you'll actually see it. A rule you can glance at is a rule you'll follow.
💡 Tip
Keep it to five or six lines. A card short enough to read in ten seconds is one you'll keep; a long policy is one you'll ignore. You can always ask an assistant to help you draft it — then edit it into words that genuinely sound like you.
Best Practices
✅ Do's
- Know and follow the rules that apply to you. Check your school's or employer's AI policy before it matters, not after.
- Disclose when it counts. A simple "I used AI to draft this, then edited and verified it" is honest and disarming.
- Own everything with your name on it. You're accountable for AI-assisted work exactly as if you'd written every word.
- Protect other people's privacy. Treat their data with the same care you want for your own.
❌ Don'ts
- Don't pass AI work off as fully your own where the rules or the situation require otherwise.
- Don't create deceptive content — fake reviews, impersonations, fabricated quotes, or misleading images.
- Don't let skills atrophy. Use AI to go further, not to stop thinking; keep your own judgment sharp.
- Don't substitute AI for a professional on serious medical, legal, financial, or safety matters.
💡 Pro Tips
- When representation or fairness matters, ask the assistant directly for a wider range of perspectives, examples, or names — it responds well to being asked.
- For anything sensitive about another person, strip identifying details before you paste, or keep it out of the tool entirely.
📓 Learning Journal
Keep a journal as you work through this guide — digital or paper. After each part, jot down:
- Key ideas you learned
- Things that clicked for you
- Questions or confusion points to revisit
- Ideas you want to try
- Your progress and how you feel about it
✍️ This part's prompt: What rules actually apply to your AI use — at school, at work, or just your own standards? Write your responsible-use card from the How To above, in your own words. Then name one time recently when you used AI (or were tempted to) where a little more disclosure, verification, or caution would have been wiser — and how you'll handle that kind of moment next time.
📝 Part Summary
🎓 Key Takeaways
- Integrity is about honesty, not abstinence: follow your school's or employer's rules, disclose AI use where it's expected, and never pass AI work off as fully your own where that's prohibited. You own everything with your name on it.
- Assistants can reproduce bias and produce misinformation; verify what you receive, and never create deceptive content or deepfakes meant to mislead.
- Guard against over-reliance and respect other people's privacy — and stay an aware, intentional user amid the real, unsettled debates about labor and the environment.
- Some things aren't for AI: high-stakes calls without a human in the loop, serious medical/legal/financial matters without a professional, and anything you can't verify but must rely on.
🎉 What You Can Now Do
You can use AI as a responsible adult, not an anxious one or a reckless one. You know how to handle integrity and disclosure, how to spot bias and misinformation, how to protect other people as well as yourself, and — most usefully — exactly when to put the assistant down and reach for a qualified human instead. That judgment is what separates a confident AI user from a careless one.
❓ Common Questions
Is it cheating to use AI for schoolwork?
It depends entirely on your school's policy, which is why your first step is to find out what's allowed. Some courses welcome AI with disclosure, some limit it, some ban it for graded work. Using it in a way the rules forbid — or passing its work off as yours where that's prohibited — is the problem, not AI itself. When unsure, ask your teacher before you use it.
Do I really have to tell people I used AI?
Not always — for a personal note or a rough draft you'll heavily rewrite, no one expects a disclosure. But where accuracy, authorship, or trust matter — graded work, client deliverables, anything published with your name — a simple, honest line about your AI use is the safe and respectful move.
Can I ask AI about a medical or legal problem at all?
Yes — to understand. It's great for plain-language explanations, background, and preparing good questions for a professional. The line is decisions: for an actual diagnosis, treatment, legal action, or money move with real consequences, see a qualified professional and treat the AI as a study aid, not the authority.
🔭 Up Next
In Part 25: Your AI Workflow — the final part — we'll pull the whole guide together into a durable personal practice: where AI fits into your week, building a library of the prompts that work for you, keeping the two habits that matter, and a warm send-off to the rest of Ray's guides and courses.
📚 Additional Resources
🌟 Keep Going
You've taken on the part many guides skip — how to use this power well, and when to set it down. That judgment will serve you long after any particular model is old news. One part remains: turning everything you've learned into a workflow that sticks. 🤖