Most coaches treat AI ethics as a single decision. Use ChatGPT or don’t. In practice, it’s dozens of small habits spread across your whole practice, from the transcription app you forgot was running to a coaching agreement you haven’t updated in years.
That’s why a one-time decision doesn’t hold up. The tools change, the terms of service change, and new AI features appear inside software you already pay for. AI ethics for coaches is less like a rule you adopt and more like a regular check-up. Here are the seven areas worth reviewing, and the questions that show you where to start.
What the ICF already expects of you
You don’t have to invent a standard from scratch. The ICF Code of Ethics, in effect since April 1, 2025, asks coaches to protect client information “directly and through any technology systems I may utilize.” Standard 2.5 names artificial intelligence explicitly in its list of those systems.
ICF’s Acceptable Use of AI and Client Protection Guidelines (revised September 2025) go further. They list appropriate uses, such as session preparation and admin. They list inappropriate ones too, including “Replacing human judgment in critical coaching decisions” and “Using AI to make clinical diagnoses or provide therapy.” And they sum up the principle in one line: “AI should inform your thinking, not replace your professional judgment.”
Even if you’re not ICF-credentialed, that’s a sensible floor. The audit below builds on it.
A seven-part audit of your AI use
Work through each area honestly. You’re not looking for a perfect score. You’re looking for the one or two areas where your habits have drifted furthest from what you’d defend to a client.
1. Do you know which tools you’re actually using?
Start with an inventory. Not just the chatbot you open on purpose, but the AI features now built into your calendar, video platform, email, and CRM. It’s easy to miss a notetaker or summary feature you never consciously switched on.
For each tool, find out two things. Does it store what you put in, and is that material used to train the company’s models? If you couldn’t explain a tool to a client in plain language, you don’t understand it well enough to use it with their information. ICF’s ethics guidance on Standard 2.5 puts the duty on the coach “to research technological tools, associated risks of the tool,” and to plan for a breach (ICF Insights and Considerations for Ethics).
2. Where does client material go?
This is the area where good intentions slip most easily. ICF’s guidelines say plainly, “Never enter personally identifiable information (PII) into open AI models,” and they name ChatGPT, Claude, and Gemini as examples of open models.
Coach Factory’s position is stricter than ICF’s minimum. We recommend keeping client material out of general-purpose AI tools entirely, even with the name removed, because a client’s story is often identifiable without it. We’ve written about where that line sits in client confidentiality and AI for coaches. A simple habit helps here. Write your session notes in your own words first, and keep them in a system you control.
3. Do clients know, and did they agree?
ICF’s guidelines ask you to “Be transparent with clients about your use of AI tools,” and suggest adding a clause to your coaching agreement. Its ethics guidance says the coach is “ultimately responsible for obtaining the informed consent of their clients” when using technology, AI included.
Good consent is specific. It names the tool, it’s given before coaching starts, and it’s written down. It also comes with a real choice. ICF advises that if a client is uncomfortable with your AI use, you should be prepared to adjust or stop. A client who says no should get exactly the same quality of coaching. If you record sessions, our post on AI notetaker consent covers what that agreement needs to include.
4. Whose insight is it?
This is the area that matters most, and it’s the easiest to rationalize away. AI belongs in your preparation and reflection. It can help you prep for a session, spot patterns in your own notes, or think through what happened afterward. It shouldn’t produce the client’s insight or make their decisions.
One useful test is to imagine dropping every AI tool tomorrow. If your coaching would get noticeably worse, AI may be doing work that belongs to you and your client. The human part of coaching is what clients are paying for, and it’s what keeps your work valuable as AI improves. We’ve made that case at length in Will AI Replace Coaches?
5. Does AI stay inside your scope?
Coaching has boundaries, and AI doesn’t respect them on its own. A chatbot will happily offer something that sounds like counseling. Make sure no AI tool in your practice is diagnosing or counseling a client, and that no automated message or chatbot can answer a client beyond what coaching covers.
Your referral plan shouldn’t depend on AI either. Knowing when to refer a client to therapy is your judgment, made with your own eyes on the situation.
6. Have you read the vendor’s terms?
Terms of service are long and dull, so it’s tempting to skip them. Still, ICF’s guidelines recommend checking how long a provider keeps data, where it’s processed, and how it handles a breach. They also suggest avoiding “free or unverified AI tools for any coaching-related work.”
You don’t need to become a privacy lawyer. Find out how long the vendor keeps your data, how to delete it, where it’s stored, and who can see it. Then recheck when the terms change. If something is unclear and the stakes are high, ask a qualified professional. (This isn’t legal advice, and the rules vary by country and state.)
7. Who checks the output, and is your approach written down?
Anything AI helps you create should be reviewed by you before a client sees it. ICF’s guidelines say to “Always review, validate, and if necessary, edit AI-generated content before using it with clients.” They also ask you to disclose when AI helped create material you share.
Finally, write your approach down. ICF recommends you “Create a written policy for your AI use” and “Periodically assess your AI usage.” A one-page statement with a review date is enough. It turns good intentions into something you can actually check against.
How to run the audit without getting overwhelmed
Seven areas can feel like a lot, especially if you’ve been using AI casually for a while. You don’t have to fix everything this week. Score yourself honestly in each area, then pay attention to your lowest one. That area is where a client is most exposed, and it’s where a small change does the most good.
Pick one change you can make in the next two weeks. Maybe it’s adding a consent clause to your agreement. Maybe it’s turning off a notetaker you don’t need. Then put a date on the calendar to run the audit again in a few months, because your tools will have changed by then.
We built a scored self-assessment that walks through all seven areas, with a tally that shows your lowest area and points you to the next step for each one.
AI Literacy & Ethics Self-Assessment for Coaches
Knowing how to use AI isn’t the same as knowing whether you’re using it well. It’s easy to start using an AI tool for notes or session prep before you’ve decided where its limits are. This is a private, scored check on your whole practice across seven areas: your tools, client da
Why this is worth your time
Clients trust you with things they haven’t told anyone else. That trust is the foundation of the work, and it’s easy to wear down without noticing, one convenient tool at a time. If you’re weighing harder situations, our guide to coaching ethics dilemmas walks through how to decide.
The good news is that none of this asks you to give up AI. It asks you to use it the way you’d want your own coach to. Carefully and openly, in service of the client. That’s a standard your clients will notice, even if they never see the audit behind it.
