Regulators Just Told Us Where AI Belongs in Mental Health. It's Not the Therapy Chair.
Four reports and rulings from the past week, from SAMHSA to the MHRA to The Lancet Psychiatry, converged on the same answer: keep clinicians in the chair, and put AI in the connective tissue around them.
Buried in a new federal report on AI in mental health services is a sentence that should reframe the debate: the strongest near-term opportunities for AI in mental health may be the least visible ones.
That is a finding, not a hedge. The report, published by the Substance Abuse and Mental Health Services Administration (SAMHSA) and co-authored by digital psychiatry researcher John Torous, reviewed the evidence across administrative, clinician-facing, and patient-facing uses of AI. The pattern it found is striking. AI's evidence base is strongest in the unglamorous places, documentation, workflow, and system-level support, and gets thinner the closer AI moves to direct therapy and crisis response. The tools that draw no headlines are the ones with evidence behind them. The tools that draw headlines are the ones we cannot yet trust.
I have spent the past year building in this space, and I read most weeks of mental-health-tech news as noise around that single signal. This week, unusually, the news pointed one direction.
Start with the regulators. The Wall Street Journal's What's News podcast ran a segment on August 23 titled "The Wild West of AI Therapy Laws," on the wave of state legislation restricting AI therapy tools. Dartmouth's Nicholas Jacobson, who builds Therabot, one of the few generative AI mental health tools with clinical trial evidence behind it, noted that these laws "don't often have exemptions for these clinical products that are effective and safe." You can read that as a complaint about blunt regulation. I read it differently. Legislatures across the country have looked at chatbots playing therapist and decided, with rare bipartisan speed, that the default answer is no. The burden of proof now sits where it belongs, on the technology rather than the patient.
The same week, a review in DIGITAL HEALTH, drawing on a symposium convened by the UK's medicines regulator, the MHRA, laid out what the next generation of oversight should look like. Its first principle: human oversight remains critical, particularly where AI informs clinical decisions or interacts directly with people seeking mental health support. Its subtler point is one builders should sit with. Safety has to extend across the entire AI lifecycle, including real-world monitoring, model drift, and adverse events, and not stop at a benchmark score at launch. A chatbot that aced its evaluation in January is not the same system in August.
SAMHSA's report adds the sharpest caveat to the "human in the loop" mantra that AI companies, mine included, like to recite. A human in the loop only counts if the human role is meaningful. A clinician rubber-stamping AI output at volume, without training, time, or accountability, is oversight theater. If we mean clinician-first, the clinician needs real authority over the tool and not a cameo in its marketing.
Meanwhile, researchers at Deakin University's Lifespan Institute published a position paper in The Lancet Psychiatry making a point the regulatory conversation keeps missing. General-purpose AI systems, ChatGPT, Claude, and Gemini among them, are often a distressed person's first point of contact, ahead of any purpose-built mental health app, and in the authors' view there are still no agreed frameworks governing it. Their proposed roadmap focuses on something I find encouraging: clarifying the optimal sequencing and integration of AI tools into clinical workflows. Not whether AI belongs in mental health care, but where in the care pathway it does the most good and the least harm.
The question has changed. Two years ago the debate was whether AI could do therapy. This week a federal agency, a national regulator, state legislatures, and a Lancet Psychiatry paper answered a different one: where AI belongs in a system of care that stays human at its core. Their answer was the infrastructure around the chair, not the chair.
That matches what clinicians have been saying. The American Psychological Association reported earlier this year that more than a third of psychologists already have patients using AI as a kind of auxiliary mental health professional, and most are far from convinced that is progress.
I think that infrastructure looks like this. A therapist sees a patient for one hour out of 168 in a week. The other 167 hours are where life happens: the sleep that collapses, the conflict that will dominate the next session. Today, almost none of that context reaches the clinician until the patient reconstructs it from memory, weeks later, in fragments. The highest-value AI in mental health is the layer that gives a qualified clinician visibility into those 167 hours, when the patient chooses to share them. In my view continuity is close to the whole mechanism of good therapy: one patient and one clinician, carried across time.
The companies that matter in this field five years from now will be the ones that made real therapists harder to replace, better informed, and present in some meaningful sense between the hours they are in the room. This week the regulators and the researchers said as much. Builders should listen.
