The Hidden Cost of Using AI as Your Therapist

People don't avoid professional mental health support because they're lazy or reckless. Research on treatment-seeking behavior points to something more nuanced: the mental health system is genuinely difficult to navigate, and the distance between recognizing a problem and getting structured help for it is larger than it should be. Waitlists stretch for months. Costs are prohibitive for many. For someone who is already overwhelmed, the friction of finding a provider, getting an appointment, and sitting with a stranger to talk about their inner life can feel like too much to take on.

AI chatbots remove that friction almost entirely. You can ask a question in plain language, get a coherent answer in seconds, and never have to make yourself vulnerable to another person in the process. That convenience is real, and it explains a lot about what a recent national poll is revealing about how people are actually managing their mental health right now.

The Numbers Behind the Trend

In March 2026, health policy organization KFF published results from a nationally representative survey of 1,343 adults on how people are using AI chatbots for health guidance. About one in three adults reported consulting AI for health advice in the past year. For those who used it specifically for mental health questions, 58% never followed up with a mental health professional afterward.

The age gap in the data is striking. Adults under 30 are roughly three times as likely as those over 50 to use AI chatbots for mental health advice. This is the generation reporting the highest rates of anxiety, depression, and emotional overwhelm. It is also the generation most likely to reach for a digital tool before considering a human one. Cost and access play a real role here: the poll found that difficulty affording care and lacking a regular provider were both cited as major reasons people turned to AI in the first place. However, the gap between having information and being able to use it is where things get complicated.

The Benefits and Limitations of AI

AI tools can be genuinely useful for mental health literacy. They can explain DBT distress tolerance techniques, describe what emotional regulation looks like in practice, and outline the difference between skills-based group programs and individual therapy. For someone who has never engaged with mental health care before, that kind of accessible language can lower the barrier to seeking help.

The limitation shows up when understanding needs to become change. An AI tool can't notice that someone's answers have shifted over several weeks, or reflect back a pattern the person themselves hasn't named yet. It can't demonstrate interpersonal effectiveness in a live exchange, hold space during silence, or offer the accountability that comes from showing up somewhere regularly. Research consistently points to the therapeutic relationship as one of the strongest predictors of treatment outcomes. A widely cited meta-analysis by Norcross and Wampold found that the quality of the relational bond between practitioner and client accounts for as much of the outcome variance as the specific treatment method used. That is not something an algorithm can replicate.

Research from Mount Sinai's Division of Data-Driven and Digital Medicine adds another layer of concern. A 2025 study benchmarking the confidence of 12 large language models across clinical questions found that worse-performing models showed paradoxically higher confidence, and that even accurate models showed minimal difference in how confidently they delivered correct versus incorrect answers. A clinician who is uncertain pauses, asks more questions, and refers out. Current AI systems don't have that calibration.

The Case for Structured Group Learning

What DBT skills groups offer isn't just information, it's a learning environment. A typical online DBT therapy session runs 60-90 minutes and moves through one of four core modules: mindfulness, distress tolerance, emotional regulation, or interpersonal effectiveness. The group format is intentional. Skills practiced alongside other people tend to stick in ways that self-directed reading doesn't, because participants aren't just absorbing concepts, they're applying them in real time with others who are doing the same work.

The relational texture of that experience is hard to replicate digitally and impossible to replicate through a chatbot. When someone in a group shares a homework example and another participant recognizes their own behavior in it, that moment of recognition is itself therapeutic. When a facilitator models holding two opposing truths at once in a live discussion, participants are watching dialectics in action, not just reading a definition of it.

For people who have been relying on AI for mental health guidance, a structured dialectical behavior therapy group represents a meaningful next step, one that moves them from reading about skills to actually practicing them with support. The accessibility gap that made AI so appealing in the first place has also narrowed considerably. Virtual DBT skills training programs have expanded significantly, making it possible to join an evidence-based group without geographic limitations or the long waitlists that come with traditional clinic-based care.

Research supports this format as more than a convenient alternative. A 2025 randomized clinical trial published in JAMA Network Open found that an online, group-based DBT skills program produced significant and sustained reductions in emotion dysregulation compared to standard care alone, with participants also showing meaningful improvements in depression, anxiety, and stress after just nine weeks of therapist-guided online sessions. The medium isn't what determines effectiveness. What matters is whether the program is evidence-based, clinician-facilitated, and built around active skill practice rather than passive content consumption.

A Signal Worth Taking Seriously

For mental health practitioners and coaches, the KFF poll is worth sitting with. Clients and prospective clients are very likely already using AI chatbots for mental health questions, not because they're disengaged from care, but because they're resourceful and the system hasn't always made it easy to be otherwise. The clinical opportunity isn't to compete with technology or dismiss it. It's to offer something more: DBT emotional regulation skills taught in context, interpersonal effectiveness practiced in actual relationship, and the kind of structured accountability that builds over time in a consistent group.

TheraHive's psychoeducational approach is designed precisely for this moment. When someone has spent months searching for coping strategies and asking chatbots about their anxiety, they often already have a framework. What they're missing is the practice and the community that turns awareness into durable change.

For anyone who has already started doing the work on their own, a structured group is where that effort starts to compound. TheraHive's Adult DBT Skills Groups are designed to take that foundation and build on it in a structured, coach-led environment alongside others doing the same work. Find out if the group is a good fit for where you are right now.

{{promo-banner-1}}