Is It Ethical for Therapists to Use AI for Marketing?

My wife Darla is a licensed mental health counselor in private practice. When we started building Ariadne, a tool that drafts weekly marketing for therapists, the first question she asked was not "will it work?" It was "is it okay for me to use this?"

That's the right first question. It's also, I have learned, the first question most therapists ask, usually with some version of I already feel weird about marketing, and now there is AI involved too.

So here is the honest answer, as plainly as I can put it.

The short version

Yes, it's ethical for a therapist to use AI to draft marketing, with guardrails. The ethical line isn't the technology. It's client information. Everything else follows from that.

Using AI to draft a Google Business post about your specialty, an introduction email to a pediatrician who might refer to you, or a blog post about what the first session is like, is ethically ordinary. It's the same activity as asking a colleague to help you word something, or hiring a copywriter, or working from a template. You're responsible for what goes out, and nothing about the tool changes that.

What wouldn't be ethical is putting anything about a client into it, or letting it say things you couldn't stand behind.

Where the real ethical lines are

Your license doesn't prohibit drafting tools. It does set rules about what your marketing can say and how you can treat client information. Those rules apply no matter who or what wrote the first draft.

1. No client information goes in. Ever. Not a name, not a session detail, not "a client I saw last week who…". Marketing is about your practice in the abstract: your specialties, your approach, who you work with in general terms, where you are, when you have openings. HIPAA governs protected health information. A marketing draft built only from information about your practice contains none.

2. No testimonials or reviews solicited from clients. The ACA Code of Ethics and the APA Ethics Code both prohibit soliciting testimonials from current clients or others vulnerable to undue influence. That includes the "can you leave us a Google review?" email that general small-business marketing tools love to send. A tool built for therapists should never offer to write one.

3. No client stories, even anonymized ones. "Composite" or "anonymized" client stories are a common marketing move, and a poor fit for clinical work. They risk identifying real people in small communities, and they imply results.

4. No outcome claims. "Overcome anxiety in eight sessions." "Save your marriage." Ethics codes require your advertising to be accurate, and outcomes in therapy aren't something you can promise. AI tools trained on general marketing copy reach for exactly these phrases, so this is where your review matters most.

5. You read and approve every word. Nothing should publish automatically. The final responsibility for anything with your name on it is yours, the same way it would be if a colleague had written it.

If a tool or a workflow respects those five, the ethics question is mostly answered.

"But will people know AI wrote it?"

This is the question underneath the ethics question for most therapists, and it's a fair one. It's less "is this allowed?" and more "will this sound like me, and will people think less of me if it doesn't?"

Two things help.

First, the ideas should come from you. The marketing that works for therapists is specific: the misconception about your specialty you correct in every first session, the reason you started working with teenagers, what you wish people knew about couples therapy before they called. A good tool starts from your own writing and your own website, not a blank prompt, so the draft carries your substance.

Second, you edit. Change the sentence that isn't how you talk. Cut the line that overpromises. The post that goes out should be one you'd have written on a good day with time to spare. That's the standard, and it's the same standard whether the first draft came from AI or from a rushed Sunday night.

Readers spot generic AI copy by its tells: breathless warmth, stock phrases like "your healing journey," hashtag soup, exclamation points. Those tells come from generic prompts, not from AI itself.

Questions to ask any AI marketing tool

Before you trust a tool with your practice's public voice, ask:

• Does it ever ask for or store client information? It shouldn't, by design, not just by policy.
• Does it post on its own? It shouldn't. You should copy and post, or at least approve each piece.
• Will it write testimonials, client stories, or review requests? For therapists, it should refuse.
• Does it make outcome claims? Look at its sample output. If it promises results, walk away.
• Can it learn your voice? If every sample sounds like the same wellness brochure, it will sound like that for you too.

Where Ariadne stands

I'd be leaving something out if I didn't say that we make one of these tools. Ariadne drafts with AI, and we say so on our homepage. We built it around the five lines above because Darla wouldn't use it otherwise: no client information ever enters the system, it never writes review requests or client stories for therapists, it avoids outcome claims, and nothing posts without the therapist reading it first.

But the guardrails matter more than the tool. If you use ChatGPT on your own, or a different product, or a copywriter, the same five lines keep your marketing inside your license.

If you want to check your own setup, our free footprint audit looks at your Google Business Profile and website and gives you a prioritized list of fixes. No signup.