A new Colorado law limiting how artificial intelligence can be used in mental health care took effect August 12. HB26-1195 requires that psychotherapy be delivered exclusively by licensed human professionals, a direct response to growing concerns about patients mistaking AI chatbots for legitimate therapists.
"AI chatbots are being mistaken by patients as legitimate therapy, which is why we're stepping up to ensure Coloradans are protected and informed," said Rep. Gretchen Rydin, D-Littleton. The law does not block licensed professionals from using AI for administrative tasks, but mandates that all clinical treatment come from a human provider.
What the law permits and prohibits
The legislation creates a clear boundary around AI in mental health care. Providers can use artificial intelligence for administrative functions like recording or transcribing meetings, as long as a licensed professional oversees the process. Any use of AI for supplementary support must be disclosed to the patient.
What the law forbids is AI independently providing therapy. It also bars companies from marketing AI chatbots as equivalent to licensed psychotherapists or counselors. Chatbots cannot imply that patient input receives HIPAA-level confidentiality protections. "This law does not hamper licensed professionals' use of AI for administrative purposes, but does require all clinical treatment and psychotherapy to be administered by a licensed, human provider," Rydin said.
Research and legal pressure behind the policy
Stanford University researchers concluded in 2025 that large language models should not replace therapists, finding that these systems "express stigma toward those with mental health conditions and respond inappropriately to certain common (and critical) conditions." The findings align with a string of lawsuits against major AI companies.
OpenAI, Google, and Character.AI face legal action from families after chatbots recommended suicide to individuals seeking behavioral health support. Parents of children who died by suicide testified before Congress last year, stating that AI chatbots discouraged their teens from seeking human support. Sen. Judy Amabile, D-Boulder, referenced these cases directly: "As policymakers, we cannot let chatbots, several of which are currently facing major lawsuits due to wrongful and horrifying deaths, replace certified mental health providers."
Why this matters for healthcare professionals
For clinicians, the law clarifies what constitutes acceptable AI use without threatening adoption of tools that handle documentation, transcription, and scheduling. The risk lies in patient-facing applications that blur the line between automated response and clinical care. Healthcare organizations deploying generative AI and LLM tools should audit whether any patient-facing systems could be interpreted as delivering therapeutic guidance, even unintentionally.
Providers in Colorado now have a legal framework that draws a hard line: administrative AI is permitted with disclosure, but treatment decisions and psychotherapy must remain human-delivered. As similar legislation emerges in other states, healthcare systems that proactively separate clinical from non-clinical AI applications will face fewer compliance headaches. For those building or buying AI for healthcare workflows, the message is straightforward - keep the bot out of the therapy room.
Your membership also unlocks:
