Therapy in the Age of AI

There’s an interesting tension happening in mental health right now, one I didn’t foresee when starting a group mental health practice. On one hand, there’s more awareness about therapy than ever before. People are talking openly about anxiety, trauma, attachment, burnout, nervous systems, boundaries. The language of emotional health has become part of everyday life in a way that genuinely feels hopeful. This is a huge cultural shift I’ve seen since entering the field.

At the same time, there’s also more confusion.

As AI enters the therapy space - through chatbots, mental health apps, self-help algorithms, and endless streams of psychological content - I find myself thinking often about what therapy actually is. Not the idea of therapy. Not the language of therapy. But the lived experience of it.

As a therapy practice in Seattle and Kirkland, we’ve always believed that healing happens in relationship. Not in perfectly optimized interventions. Not in quick fixes, or in consuming more information. But in the slow, often uncomfortable experience of being known over time.

That can feel countercultural right now.

We live in a moment that values speed, efficiency, and immediate answers. AI is remarkably good at delivering those things. It can summarize, suggest, reassure, organize, and even mimic empathy in ways that are impressive. I’m not anti-AI. In many ways, I think it will become an incredibly useful support tool in healthcare and mental health, and help many people access support and information who may otherwise not be able to.

But support is not the same thing as relationship.

One of my concerns is that people may begin mistaking emotional fluency for emotional intimacy. We can now access endless content explaining why we feel the way we do. We can ask AI to analyze our relationships, identify our attachment style, or generate coping skills in seconds. And sometimes that information is genuinely helpful!

But insight alone rarely transforms us.

Real therapy as I see it - especially long-term relational work - asks something much deeper. It asks us to stay. To tolerate vulnerability. To let another person witness the parts of ourselves we usually organize, hide, or intellectualize away. It asks us to practice trust, repair, honesty, grief, dependence, autonomy, and integration in real time with another human being.

That process cannot be automated.

Ironically, I think the rise of AI may make deeply relational practices even more important. As the world becomes more efficient, people will continue hungering for spaces that feel human. Spaces where they are not optimized, categorized, or managed - but actually encountered in real time.

Running a boutique relational therapy practice in this moment means holding onto that belief very intentionally. We are not trying to scale emotional care into something frictionless. We are trying to protect the depth, slowness, and humanity that healing and compassion often requires.

I suspect the practices that endure will not simply be the most efficient ones, but the ones that help people feel deeply seen in a world increasingly designed for speed.