Let's be honest: therapists are skeptical of AI. And they should be.
When you work with people's deepest fears, their trauma, their most vulnerable moments, the idea of a machine listening in doesn't feel neutral. It feels risky.
One therapist told us that about 75% of her clients would not be comfortable with AI in session. "The clients that I serve, people of color, the queer community, they are a little bit wary of having AI in session listening in."
Another clinician hasn't even tried the AI features her EHR offers, partly because many of her clients are physicians who are "really concerned about data security and privacy." She worries that even asking for consent might be a turn-off.
And a third was more blunt about the stakes: "Some of this stuff, it would be really harmful to the client if it came out."
So yeah. Skepticism is earned.
But here's the thing: every therapist we talked to is already using AI. They're just using it carefully.
They use ChatGPT to organize intake notes, to find a worksheet they remember from years ago, to draft website copy, and to brainstorm interventions when a session feels stuck. One therapist put it this way: "Here's the things that we've already tried. What's something we've missed?"
What they want isn't AI that replaces them. It's AI that supports their judgment.
As one clinician told us: "You still have to use clinical judgment. I think it should be a disclaimer." And another: "I just don't want to be replaced by AI."
Helpful is good. Creepy is not.
The message is clear. Transparent is essential.
At Harbor, AI is opt-in and explainable. Clients know when it's active. Therapists control what it sees. And it never makes decisions. It just offers support.
Because trust isn't a feature you can bolt on. It has to be built in from the start.