Therapy robots are already here

Ryan McGuire / Gratisography / Used under licenseTherapists often fear manualized treatments in psychotherapy. If the therapy process is boiled down to a script, the fear goes, the actual therapist becomes interchangeable with anyone else following the same script. Taken to its logical end, if therapy is just a set of manualized techniques, we could easily be replaced by robots.

So, that’s happening. The Guardian published a story recently about a therapy chatbot powered by artificial intelligence. Not only does the bot learn from its mistakes on a broad level, it also learns on an individual client level what responses work well, and what don’t. The robot is programmed to notify humans in the event that any crisis issues emerge, but is seen as a strong frontline support for many clients.

If you learned that therapy is simply a set of theory-based techniques, you have reason to worry. And that’s largely how therapy is taught these days: As a set of theory-based techniques. With an ever-expanding list of manualized treatments to teach, graduate programs are getting longer and thus more expensive. Student loan debt has become a heavy burden for new therapists. And there’s no data to suggest that those longer and more expensive programs make therapists any better. Overall outcomes in therapy aren’t improving.

Those graduate programs, it seems, are teaching the wrong things.

The syringe versus the medicine inside it

Let’s suppose, for a moment, that you were about to enter medical school. On that first orientation day, a professor tells you that you will spend most of your time in school learning about syringes – fat ones, skinny ones, short ones, long ones, old ones, new ones. There are a lot of different kinds of syringes! Mind you, the new ones don’t work any better than the old ones, but each one has to be handled a little differently, so you’re going to be learning them all.

Wouldn’t you think twice about the whole educational process? Wouldn’t you wonder why you spending so much time learning about varieties of things that don’t actually make much difference in getting medicine into the vein? Wouldn’t you wonder why you weren’t learning more about the actual medicine that goes into the body?

But that’s exactly how we teach therapists. The therapeutic model is our syringe. It is a delivery device for what is curative, not the actual thing that is curative. And what is actually curative? The relationship between therapist and client. Or if you prefer to use the more formal research term for the part of the relationship that the therapist can control: therapist way of being. And therapist way of being is, in many programs, only taught indirectly — if it is taught at all.

This is not, by and large, the fault of those programs. They respond to pressures from accreditors, who themselves are often responding to pressures from employers, who are demanding that therapists be trained in the kinds of therapy that insurance companies and other large health care systems will pay for. Those payors understandably want accountability. They have every right to know that their money is being spent as efficiently as possible. And so therapy programs train their students in manualized approaches that bear the “empirically supported” stamp, without attending to the fact that that doesn’t mean anything. And researchers continue trying to build a better therapeutic mousetrap, crafting models full of intervention and missing a core element of human connectedness.

It is probably true that, no matter how much we may wax poetic about some metaphysical connection that happens between therapist and client, this too can eventually be meaningfully replicated by robots. So in a sense, all of our jobs are in eventual danger to robots, I suppose. But the ones who learned (incorrectly, by the way) that therapy is merely technique, that manualized treatments are best for our clients — those are the ones who will be replaced first.