There’s a lot being written about the ethics of people using AI for therapy, and I’m not going to add to that pile. Whether an AI gives people the right kind of support is a whole argument of its own, and others are having it. What I want to talk about is what’s actually happening, and what it means for your practice.
Because people are already doing it. Right now, in large numbers, people are opening ChatGPT and talking through what’s going on in their heads. And here’s the bit that makes a lot of therapists wince: for a certain kind of work, it’s actually quite good.
What AI does well, honestly
Think about what a lot of talk therapy actually consists of. A conversation with someone who listens, asks good questions, and helps you think through what’s going on. Someone who reflects your thinking back at you in a way that raises your self-awareness and deepens your insight. That’s real work, and for decades the only way to get it was to sit across from a trained professional.
AI can now do a version of that. Someone who isn’t in crisis, who just wants to understand themselves better, can talk things through with an AI at 2am, for free, for as long as they like, and come away with insights that feel meaningful. Whether it’s the right kind of support is, as I said, a different argument. But the person doing it doesn’t think they’re in a philosophy debate. They think they’re getting somewhere.
Now for the scary bit: if the core of what you offer is insight through conversation, AI represents a real threat to your livelihood. Not because it’s better than you. Because it’s free, it’s available, and for the self-understanding end of the work, it’s good enough that people are choosing it. I can see why a lot of therapists are going to lose out to that, and I don’t say it with any glee.
Why HTC sits differently
You can’t do Head Trash Clearance with an AI. Let me explain.
AI doesn’t know the method. HTC isn’t in the training data in any way an AI can use. I’ve been trying to train AI on the method for years, properly, with everything I know, and it still doesn’t produce the mantras correctly. The hard part of this work is identifying the charge: knowing which theme is live, which word carries it, what to clear and in what order. This has to be done by a human. And it’s the skill you build on the training. It’s not easy to grasp. And I say that as someone who’s trained therapists for months who still wrestle with it. It’s why the training takes eight months rather than a weekend.
The work is energetic. HTC is energy work, and part of what you learn as a practitioner is energetic testing: tuning in to your client’s system to read what’s there. Whatever AI ends up being able to do, it isn’t in the room with a human energy system, and I don’t see how it ever gets there. That ability to tune in energetically to a client is where the real value of a session lives, and it belongs to the practitioner.
The two questions clients always come back with. I’ve built AI tools for my practitioners and I provide resources to my clients, so my clients are as equipped as anyone on the planet to run this method on their own. And they still come back, every time, with the same two questions: is the thing actually cleared, and what do I need to clear next? Answering those takes the testing, the pattern recognition, and the fifteen-plus years of knowing where charge hides. That’s the practitioner’s job. It’s the part I can’t hand to an AI. I know because I’ve tried.
The part that might surprise you
Here’s what I’m not going to claim: that clients using AI between sessions is a problem. In my practice it’s the opposite, and I encourage it.
One of my longest-standing clients has spent years learning this method with me, been mentored through it, and invested heavily in the work we’ve done together. She’s now got her AI supporting her between our sessions, running the pieces she knows how to run. It isn’t replacing me. She’s doing it so the work keeps moving when we’re not together, and then she brings it back into session, where the deeper work happens. That’s not a threat to my practice. That’s my delivery model working exactly as designed.
Because here’s the thing about HTC: I never gatekept the self-serve end of it. The method is in my book for a few pounds. Clients doing change work in their own time, between sessions, is the whole point of how this is delivered: the session sets up and supervises the work, and the client runs it in their week. Read: how the HTC delivery model works. If AI helps a client do more of their between-session work, the work compounds faster, and the practitioner still holds everything that makes it land: the protocol design, the testing, the judgement about what’s cleared and what’s next.
So the same wave hits the two models very differently. AI erodes a practice built on insight through conversation, because it can imitate the product. It strengthens a practice built on the protocol model, because it enriches the layer the client was always meant to own, while the practitioner’s layer stays out of reach.
What this means if you’re weighing up your future
I’d put it this way. The question isn’t whether AI changes the therapy landscape. It’s already changing it, and the people it affects most haven’t felt it yet. The question is whether the thing you’re trained in sits on the side of the line AI can imitate, or the side it can’t.
Insight through conversation: AI can imitate that today. Clearing charge from a human energy system, knowing what’s cleared and what to hunt down next: it can’t, and everything I’ve seen in years of trying tells me it won’t.
That’s worth thinking about when you choose what to train in next.
Curious about adding HTC to your toolkit? Explore HTC Practitioner Training.