When the Client Has Already Asked AI: Rethinking the Therapeutic Relationship in the Age of Chatbots
A client arrives at a session and says, “I asked a Chatbot about this before I came here.”
For mental health professionals, this may evoke a complicated mixture of reactions: curiosity, concern, scepticism, irritation, or perhaps an uncomfortable sense of professional displacement. Artificial intelligence is increasingly entering not only the everyday lives of clients, but also the interpretive space traditionally occupied by therapists, helping people name emotions, make sense of relationships, formulate possible explanations for their distress, and search for reassurance.
This is no longer a hypothetical issue. The American Psychological Association’s 2026 Chatbots and Mental Health Survey found that 77% of surveyed psychologists had spoken with patients who had used AI for support, engagement, or other reasons. Nearly two in five reported patients using AI to self-diagnose, while 35% said patients were using AI as an additional mental-health professional.
The question, therefore, is not simply whether clients should be using AI. A more clinically useful question may be: what is the client using it for?
AI as a meaning-making tool
People have always searched for ways to understand experiences that feel confusing, frightening, or difficult to articulate. What is different now is the immediacy of the response. A person can describe an argument with a partner, a recurring thought, a difficult childhood memory, or a vague sense of emptiness and receive an organised narrative within seconds.
That response can be reassuring. It can also be misleading.
AI can provide language for experiences that a client has struggled to name. It may help someone formulate questions before a therapy appointment or organise thoughts they have difficulty expressing. The APA similarly notes that AI can potentially help clients organise their thoughts, generate ideas, and practise therapeutic exercises learned through treatment.
But the fluency of an answer can easily be mistaken for the validity of an answer. AI systems can sound confident while being inaccurate, and they may respond in ways that reinforce a user's existing interpretation rather than meaningfully challenge it. This creates a particularly important clinical tension: the experience of feeling understood is not necessarily equivalent to being accurately understood
What does the therapist experience?
The arrival of AI also makes the therapist part of the equation.
When a client says, “The AI told me I have an attachment issue,” the therapist may feel an immediate pull to correct the information. When a client says, “I talk to it when I'm lonely,” the therapist may wonder what the chatbot is providing that the therapeutic relationship is not.
These reactions are worth noticing.
AI is not only changing client behaviour; it may also expose therapists' assumptions about expertise, authority, and the nature of therapeutic work. If part of our professional identity is built around being the person who understands, interprets, and explains, what happens when clients arrive having already consulted another source that appears endlessly available and endlessly responsive?
Perhaps this is precisely where reflection becomes important.
The most productive clinical stance may not be to compete with AI, nor to accept it uncritically.
Instead, its presence can become part of the therapeutic conversation.
What was the client hoping to find?
Why did they turn to AI at that particular moment?
Did they want information, validation, certainty, companionship, or simply somewhere to put a thought into words?
Did the interaction make them feel more capable of understanding themselves, or more dependent on an external answer?
These questions shift the focus from the technology itself to the psychological function it serves.
That distinction matters. A client's use of AI may tell us something about their relationship with uncertainty, reassurance, autonomy, help-seeking, or self-understanding. In this sense, the chatbot is not necessarily outside the therapeutic process. It may already be part of the client's meaning-making environment.
The limits matter too
Curiosity, however, cannot substitute for critical evaluation. Psychologists surveyed by the APA expressed substantial concerns about privacy, bias, inaccurate outputs, reinforcement of harmful or distorted thinking, and the possibility of dependency.
The task, then, is neither to dismiss AI nor to surrender clinical judgement to it.
It is to remain attentive to what technology can offer and what it cannot.
The therapeutic relationship is not valuable simply because a human is present instead of a machine. It is valuable because therapy involves context, accountability, relational understanding, ethical responsibility, uncertainty, and the ongoing capacity to reconsider what we think we know.
Perhaps the emerging question for therapists is therefore not, “Will AI replace us?”
It is a more demanding one:
When our clients can ask a machine almost anything, what makes the therapeutic conversation worth having with another human being?
That question may be less about defending therapy from technology and more about clarifying what we believe therapy is for.
Author: Jahanvi ( Counselling psychologist)