Can AI Really Hold Us? Reflections on Therapy, ChatGPT, and Being Truly Seen

Can AI Really Hold Us? Reflections on Therapy, ChatGPT, and Being Truly Seen

Can AI Really Hold Us? Reflections on Therapy, ChatGPT, and Being Truly Seen

“The act of revealing oneself fully to another and still being accepted may be the major vehicle of therapeutic help.”
— Irvin D. Yalom, The Gift of Therapy, Chapter 4

There is a growing trend that I am hearing more and more about in my work: clients turning to AI, and particularly ChatGPT, to help them make sense of how they are feeling, to organise confusing thoughts, or to seek validation when they feel misunderstood. In a world that is fast paced, pressured, and often overwhelming, this makes a lot of sense. AI is available instantly. It doesn’t get tired. It doesn’t interrupt. It offers words when we don’t have any.

I use ChatGPT myself. I find it an incredibly helpful and resourceful tool, especially when I need quick access to information, a shortcut to an answer, or a deeper dive into a topic that would otherwise take time and effort to research. In many ways, it fits perfectly into the reality of modern life. When time is limited, AI can feel like a lifeline.

But when it comes to therapy, or something that begins to resemble it, I find myself wondering: how helpful is this really?

Yalom’s words from The Gift of Therapy feel particularly relevant here. He speaks about the power of revealing oneself fully to another and still being accepted. Not analysed. Not corrected. Not reframed immediately. But accepted. And crucially, accepted by another human being.

If we are asking AI to help us understand how we feel, how we think, how we behave, and how others behave towards us, what exactly is happening in that interaction? Is that an intervention? Is it therapeutic change? Or is it something else entirely?

AI does not see us. It does not sit opposite us and witness our struggle. It cannot feel the emotional weight of our story or sense the wounds that have been shaped through a lifetime of relationships. It works only with what we give it. And that matters.

What we put into AI is inevitably a partial story. A snapshot. A moment in time filtered through our own perception, mood, and pain. The response we receive is therefore also partial. It can sound empathic, validating, even wise. But it is still a biased exchange: a biased input receiving a biased output, without the possibility of being gently challenged, expanded, or held within a relational context.

This is where Imago Relationship Therapy offers something fundamentally different. Imago is built on the understanding that healing happens in a relationship. Not just any relationship, but in the intimate partnership we have chosen, and where we ourselves have been chosen in return. The relationship becomes the space where old wounds are activated, yes, but also where they can be healed.

In Imago, the work is not about being right or proving reality. It is about being seen, heard, and validated by our partner, even when their experience is different from our own. Especially when it is different.

Something I am noticing more frequently is couples using AI to check whether their “reality” is real, particularly when their partner is telling them it isn’t. And this is deeply painful territory. When someone is told that their feelings, perceptions, or memories are “not real”, the injury is not just emotional; it is relational.

From an Imago perspective, the simple truth is this: both realities are real. Each partner’s internal experience is valid, even when those experiences don’t match. One partner’s truth does not cancel out the others. Healing begins not when we agree, but when we become curious.

Imago teaches us how to slow down, listen deeply, and reflect our partner’s world back to them accurately, without judgement or correction. Through dialogue, we learn to create empathy for the person sitting in front of us, a person shaped by their history, their wounds, their unmet needs, and their longing for connection. This is not something AI can facilitate between two people. It cannot sit in the space between partners and help them encounter one another with safety and intention.

At best, AI can validate one side of a story. At worst, it can harden positions and deepen disconnection.

And then there is an irony that I can’t ignore. As I write this, I am aware that I could easily paste these words into ChatGPT and ask it to rewrite them, so they sound “better”, clearer, more acceptable, more polished. Perhaps that is exactly what many people are seeking when they turn to AI: help to present themselves in a way that feels safer, more coherent, more socially acceptable.

But Imago invites something braver than that. It asks us to bring not just the polished version of ourselves into relationship, but the vulnerable, unfinished, messy parts too. The parts that fear rejection. The parts that learned early on that being fully seen was not always safe.

Therapy, and particularly relational therapy, is not about finding the right words. It is about creating the conditions where truth can emerge and still be met with acceptance. AI can be a useful tool. It can inform, organise, reflect, and even soothe in the moment. But it cannot replace the profound healing that comes from being fully revealed and fully accepted by another human being.

Especially by the person we chose.
And the person who chose us.

And perhaps that is the question we need to keep asking:
Are we looking for answers, or are we longing to be seen?

Because those are not the same thing.