Colorado's AI Psychotherapy Rules Take Effect August 12

Colorado's mental health landscape is on the cusp of a significant shift, with new restrictions on artificial intelligence in psychotherapy services set to take effect on August 12, 2026. This upcoming deadline means that mental health professionals and entities across the state have just days to ensure their practices align with the state's commitment to human-centered care.

What's changing

The "Act Concerning restrictions on the use of artificial intelligence relating to psychotherapy services" (HB26-119) was signed into law by the Governor on June 3, 2026, and its core provisions will become enforceable starting August 12, 2026. This landmark legislation, specific to Colorado, aims to safeguard the integrity of therapeutic relationships by clearly defining the boundaries for AI integration in mental healthcare.

At its heart, the Act explicitly prohibits AI systems from engaging in direct therapeutic communication with clients. This means that AI cannot independently conduct therapy sessions, offer counseling, or serve as a primary point of contact for clients seeking mental health support. The intent is clear: the therapeutic process must remain a human-to-human interaction, preserving empathy, nuance, and professional judgment.

Furthermore, the Act restricts AI from generating unsupervised treatment recommendations or attempting to detect a client's emotions or mental states without direct human professional oversight. While AI can process vast amounts of data, the critical interpretation, diagnostic formulation, and treatment planning must ultimately reside with a qualified, licensed mental health professional. Any AI-generated insights must be thoroughly reviewed, validated, and explicitly approved by a human practitioner before being applied or communicated to a client.

However, the legislation isn't a blanket ban on AI. It acknowledges the potential for AI to serve as a valuable administrative and supplementary tool. Practitioners can still leverage AI for tasks such as scheduling appointments, managing billing, or transcribing sessions. Crucially, even for these supportive functions, the Act introduces a vital requirement: if AI is used to record or transcribe sessions, practitioners must obtain clear, written informed consent from the client. This provision ensures transparency and client autonomy, preventing the surreptitious use of AI in sensitive therapeutic environments.

The human professional always retains full responsibility for any AI interactions or outputs. This means that accountability for the quality, ethics, and legality of services provided, even those facilitated by AI, rests squarely on the shoulders of the licensed practitioner.

Who is affected

This Colorado Act casts a wide net, impacting virtually anyone involved in providing psychotherapy services within the state. It applies to all licensed mental health professionals, including but not limited to:

Licensed psychologists
Licensed professional counselors
Licensed clinical social workers
Licensed marriage and family therapists
Licensed addiction counselors
Candidates or unlicensed psychotherapists working under supervision

Beyond individual practitioners, any entity offering psychotherapy services in Colorado must also adhere to these new regulations. This includes private practices, clinics, hospitals, and telehealth platforms operating within the state's jurisdiction.

Violations of this Act carry significant consequences. For licensed professionals, non-compliance can lead to disciplinary actions from their respective licensing boards. These actions may include substantial fines, potentially up to $5,000 per violation, and could even result in the suspension or revocation of their professional license. Such penalties underscore the seriousness with which Colorado views the ethical and responsible use of AI in mental healthcare.

Moreover, the unlawful provision or advertising of AI-driven psychotherapy services that contravene the Act can be deemed an unfair trade practice under the Colorado Consumer Protection Act. This broader legal framework allows the state's Department of Law to pursue civil penalties, which can reach up to $20,000 per violation. This dual enforcement mechanism highlights the state's commitment to both professional standards and consumer protection.

A key pitfall for practitioners could be overlooking the informed consent requirement for administrative AI uses, such as transcription. While seemingly innocuous, failing to secure explicit written consent for such activities could lead to penalties, demonstrating the need for a thorough review of all AI-related workflows.

Three things to do this week

With the August 12, 2026, effective date just days away, mental health professionals and organizations in Colorado should take immediate action to ensure compliance:

Review and Restrict Direct AI Engagement: Conduct an urgent audit of all AI tools and platforms currently in use or under consideration. Ensure that no AI system is configured or permitted to directly engage in therapeutic communication with clients. This means AI should not lead conversations, provide direct advice, or act as a primary therapeutic agent.
Implement Human Oversight for AI Recommendations: Establish clear protocols requiring that any AI-generated therapeutic recommendations, treatment plans, or insights are thoroughly reviewed, validated, and explicitly approved by a regulated human professional before being presented to a client or integrated into their care plan. AI should serve as a support tool, not a decision-maker.
Cease Unsupervised AI Emotion Detection: Immediately discontinue the use of any AI systems designed to detect or interpret clients' emotions or mental states without direct, real-time human professional oversight and validation. The Act specifically prohibits this unsupervised application to prevent misinterpretation and ensure clinical accuracy.
Secure Written Informed Consent for AI Transcription/Recording: For any administrative uses of AI, such as transcribing or recording therapy sessions, develop and implement a robust process for obtaining clear, written informed consent from clients. This consent must explicitly inform clients about the AI's role and how their data will be handled.

Related context

Colorado's Act concerning AI in psychotherapy services is part of a broader legislative trend addressing the responsible integration of artificial intelligence across various sectors. This specific Act complements other regulatory efforts, such as Colorado's own "Act Concerning the use of artificial intelligence in health care" (/regulations/RAI-US-CO-HB26113-2026), which likely addresses AI in healthcare more generally.

Beyond Colorado, other states are also grappling with similar challenges. For instance, Illinois introduced the "Wellness and Oversight for Psychological Resources Act" (/regulations/RAI-US-IL-HB18060-2025), and Maine has proposed "An Act to Regulate the Use of Artificial Intelligence in Providing Certain Mental Health Services" (/regulations/RAI-US-ME-LD20820-2026). These parallel initiatives underscore the growing recognition of the need for clear guidelines to ensure that AI technologies enhance, rather than compromise, the quality and ethics of mental health services.

Note: this article was drafted by AI - Google Gemini