When the patient arrives at therapy with their chatbot: the APA flags a dynamic the profession cannot ignore
🕒 Published on AI Momentum: June 30, 2026 · 03:40
The American Psychological Association puts a name to something many therapists already sense: their patients process their problems with AI before—or instead of—coming to the consultation. The phenomenon redefines the therapist's role and raises questions the profession still cannot answer.
By Momentum IA · June 29, 2026.
The available material for this piece is limited to the headline and the institutional reference —the full content of the APA report could not be downloaded—, so let's be direct: a little that's honest is worth more than a lot that's fabricated.
What is on the table is the fact itself: the American Psychological Association, one of the most influential institutions in mental health globally, has published in 2026 a specific report on chatbots and mental health whose headline reads 'Patients are bringing AI to therapy.' That the APA frames it this way —not 'AI comes to the clinic' but 'patients are bringing AI to therapy'— is already a highly significant data point. The agent of change is not the healthcare system or the therapist: it is the patient.
As industry context, this responds to a well-documented trend over the past two years: millions of people have begun using general-purpose language models (ChatGPT, Claude, Gemini) or dedicated mental health apps (Woebot, Wysa, among others) to process anxiety, grief, relational conflicts, or simply to be heard at three in the morning when no professional is available. This use is neither marginal nor anecdotal: consumer surveys in the U.S. and Europe show that a growing share of those who do have access to conventional therapy combine it with regular AI interactions.
This places the therapist in an unprecedented position. Before, the waiting room was the boundary of what the patient had processed alone. Now, that patient may arrive having had fifty exchanges with a model that has helped them —or confused them, or reinforced their biases, or simply kept them company— before sitting down on the couch. The therapist does not know what happened in those conversations, what interpretive frameworks the patient has absorbed, nor what kind of emotional bond they have developed with the machine.
Our reading is this: the irruption of AI into the therapeutic space is not a simple threat to the profession, but it is a real complication and, in the short term, very poorly managed. Professional associations, ethical protocols, and therapist training lag far behind the actual behavior of their patients. That the APA devotes a dedicated report to this in 2026 is a sign that the institution recognizes the urgency, but the recognition comes late relative to the phenomenon.
In the long term, however, the direction may be positive: the brutal shortage of mental health professionals —there are countries where the wait to see a psychologist exceeds a year— makes AI as a support tool between sessions, or as a first point of contact, not only understandable but necessary. The problem is not that patients use AI; the problem is that they do so without clinical guidance, without clear referral protocols, and without their therapists being prepared to integrate that information into the process. That is where the profession has urgent work to do.