As a therapist, I have noticed a growing number of clients using artificial intelligence (AI) before reaching out to a licensed psychotherapist. This trend — where an AI chatbot may be the first place humans ask questions about anxiety, relationships, trauma, depression, or other emotional concerns — is growing.
A 2026 study published in JAMA Pediatrics found that 19.2% of adolescents and young adults surveyed in 2025 had used an AI chatbot for mental health advice, compared with 13.1% in 2024. Among those who used chatbots for mental health advice, most described the experience as at least somewhat helpful.
When I first noticed this trend in my practice, I felt uneasy. How accurate is the information clients receive? What about privacy and safety? And could AI eventually replace therapists?
From Unease to Acceptance
Over time, I realized I needed a different approach: acceptance. AI is here to stay. Rather than viewing it only as a dubious threat, therapists may need to learn how to work with clients who are using it.
AI is increasingly becoming part of clients’ emotional and informational support systems. Some enter therapy after asking a chatbot about symptoms, relationships, coping strategies, or possible diagnoses. In that sense, they are bringing a new, nonhuman part of their support network into therapy.
What AI Can’t Replace
I do not believe that AI can currently replace the human therapeutic relationship. Information alone is rarely enough to heal emotional wounds. Psychotherapy with a human offers something different: empathy, attunement, trust, presence, relational repair, and the experience of being understood by another person.
At the same time, AI can have value. Chatbots can provide accessible information, help clients organize their thoughts, and introduce psychological concepts. Used thoughtfully, AI may augment psychotherapy rather than replace it.
The question for me is no longer, “Will AI replace therapists?” It is, “How can therapists help clients use AI thoughtfully while preserving what is uniquely human about psychotherapy?”
1. Get Curious About Their AI Use
I begin with curiosity. What are they asking the chatbot? What advice have they received? What has been helpful, confusing, or incomplete?
Their use of AI may reflect a need for reassurance, information, validation, or immediate support. Understanding how they use AI can also help me better understand what they are seeking emotionally.
2. Stay Nonjudgmental
Therapists should approach clients’ AI use with curiosity rather than criticism.
Clients may turn to chatbots because they want information, feel uncomfortable asking another person, or are trying to manage distress. If they feel judged, they may stop discussing their AI use or disengage from therapy.
We do not have to agree with everything AI tells a client in order to remain curious about why the information matters to them.
3. Name the Disagreements Openly
I am honest with clients that my clinical perspective may sometimes differ from what a chatbot has suggested.
Rather than treating this as a competition between therapist and AI, we can examine the information together. Does it fit the client’s circumstances? Is important context missing? Is general information being interpreted as personalized clinical advice?
Ultimately, neither the therapist nor AI should make decisions for the client. Our role is to help clients think clearly and make informed choices.
4. Compare What Therapy and AI Each Offer
Human psychotherapy provides empathy, presence, emotional attunement, eye contact, and the ability to notice subtle changes in emotions, mood, behavior, and tone. It also offers an ongoing relationship in which trust, vulnerability, challenge, and repair can develop over time.
AI offers different advantages. It can provide information quickly, explain psychological concepts, and help clients generate questions and ideas.
Clients should also understand that the confidentiality protections associated with psychotherapy may not apply in the same way to information entered into AI platforms.
5. Set Clear Boundaries
In my practice, I welcome clients to discuss information they have received from AI, but I ask them not to use chatbots during therapy sessions.
I also generally discourage clients from sending me lengthy AI transcripts or reports. Instead, I ask them to verbalize what stood out in their own words: How did they understand what AI said to them? What affected them? What did they agree or disagree with?
My therapeutic relationship is not with the chatbot. It is with the client and their experience of what the chatbot told them.
Information Isn’t the Same as Change
AI will likely become increasingly integrated into how people seek mental health information. Therapists will need to continue thinking about accuracy, privacy, safety, and appropriate boundaries.
But AI may also remind us of something fundamental: People do not come to therapy only because they lack information.
A person may already know that a relationship is unhealthy, that childhood experiences continue to affect them, or that they need stronger boundaries. Knowing something intellectually does not necessarily make emotional change easier.
Psychotherapy provides a relationship in which those patterns can be understood, experienced, challenged, and gradually changed.
AI may become a useful companion in that process. But the heart of psychotherapy remains relational.
Perhaps the task ahead is not to defend therapy from AI, but to become clearer about what each can offer and to help clients use both thoughtfully.
