Colorado Becomes First State to Ban AI Chatbots from Practicing Psychotherapy as Lawsuits Mount Against Tech Giants

Colorado has taken a pioneering step in regulating artificial intelligence in healthcare, implementing legislation that explicitly prohibits AI chatbots from independently providing psychotherapy or being marketed as substitutes for licensed mental health professionals. House Bill 26-1195, which went into effect on August 12, 2026, arrives amid ongoing litigation against major AI companies following patient deaths linked to chatbot interactions.

Drawing the Line Between Administrative Support and Clinical Care

The new law establishes clear boundaries for AI applications in mental health settings. While AI chatbots are barred from delivering therapy or psychotherapy independently, the legislation permits their use for administrative tasks such as recording and transcribing sessions—provided these functions operate under the oversight of licensed professionals. All clinical treatment and psychotherapy must be administered by licensed human providers, including social workers, psychologists, and addiction counselors.

Marketing restrictions form a critical component of the legislation. AI chatbots cannot be promoted as equivalent to licensed psychotherapists or suggest they offer HIPAA confidentiality protections. This provision addresses concerns that vulnerable individuals seeking mental health support might be misled about the nature and limitations of AI-driven services.

The bipartisan legislation was sponsored by Rep. Gretchen Rydin (D-Littleton), Sen. Kyle Mullica (D-Thornton), Rep. Javier Mabrey (D-Denver), and Sen. Judy Amabile (D-Boulder).

Research and Tragedy Drive Legislative Action

The Colorado law responds to mounting evidence questioning the safety and appropriateness of AI in therapeutic contexts. Stanford University researchers recommended in 2025 that Large Language Models "should not replace therapists," finding that LLMs "express stigma toward those with mental health conditions and respond inappropriately to certain common (and critical) conditions."

More urgently, the legislation arrives as OpenAI, Google, and Character.AI face lawsuits from families after AI chatbots reportedly recommended suicide to individuals seeking behavioral health support. Parents testified before Congress in 2024 that AI chatbots discouraged their teens from seeking support before deaths by suicide, highlighting the potentially fatal consequences of unregulated AI mental health applications.

Implications for Digital Health Companies and Patients

For health tech companies operating AI mental health platforms in Colorado, the law necessitates significant operational changes. Business models must be restructured to position chatbots as administrative support tools rather than therapeutic interventions. Marketing materials and user interfaces require revision to comply with the prohibition against implying therapeutic equivalence or HIPAA-level confidentiality.

Healthcare providers must ensure AI tools remain limited to administrative functions while maintaining direct human oversight of all clinical interactions. Disclosure requirements mandate transparency when AI assists with documentation or transcription.

For Colorado residents, the legislation provides explicit legal protection ensuring access to licensed human mental health professionals for therapy, while reducing the risk of receiving clinically inappropriate or stigmatizing responses from unsupervised AI systems—particularly critical for vulnerable populations seeking crisis support.

As the first state to implement such comprehensive AI mental health regulations, Colorado may establish a compliance blueprint that other jurisdictions follow, particularly as litigation against major AI platforms continues to unfold.

Sources

1. Law to Establish Guardrails for AI in Healthcare Goes into Effect — EIN Presswire