People now tell AI systems even their most private thoughts.

Because where people take shelter has changed.

People used to entrust their pain to a friend, a notebook, a line of prayer, a therapist, sometimes to the silence of the night that answers to no one. Today some of those lines are typed into a screen. People now open their loneliness, their shame, the knots they cannot untie not only to another person but to a screen that answers back.

That is why, for me, AI is not only a technical advance. It is the mood of an era. One of the new mirrors in which we look at ourselves.

In a counseling room in Istanbul I have watched for years how anxiety tightens in the body and how loneliness thins a person's voice. The search for belonging can carry someone even out of their own story. The same sentences are now typed into a chatbot or a personalized AI tool, in Komotini, in Istanbul, in Dublin.

People have not changed. Only the place where they tell their troubles has.

So the relationship between mental health and AI cannot be understood in the language of model capacity, efficiency, or technological progress. Any system that touches a person where they are fragile, whether in the therapy room or behind a screen, carries a psychological responsibility. Because the question is not what is possible; the question is what the possible does to a person.

My work stands exactly on this threshold. How can generative AI be used in mental health safely and accountably, with human dignity kept in view? As a system comes to know someone better, does it truly draw closer to them, or only touch the weak spot more skillfully? When does personalization become support, and when does it turn into a pull that feeds a cycle of attachment and surrender?

My concern is not to put AI in the place of therapy or of human relationships. Such a claim would misread both the person and the technology. My concern begins somewhere more basic.

We cannot assume that a technology which touches the human mind is innocent.

We cannot settle for a system that enters a person's privacy merely because it works.

We cannot exempt a tool that answers human pain from clinical responsibility.

This is why I study AI. Because the work of a mental health expert does not end with waiting in the therapy room. Wherever people carry their troubles, whatever new door they leave their hope at, clinical attention has to go there too. Someone typing into a screen in Komotini deserves that same attention, even if they never walk into the room.

What I do is build a language of safety between psychology's long understanding of the person and the new obligations of the age of AI. That language has to carry clinical responsibility into the engineering decision itself.

Because the future will also be built by those who ask, in time, what these systems do to the human spirit.

What happens when a person is left alone with the answer? Asking that in time is the name of the work I do.