A nationwide survey commissioned by ICANotes, a provider of behavioural health electronic health record systems, indicates that many mental health professionals in the United States believe artificial intelligence tools used for emotional support may be delaying patients from seeking qualified care.
The study, carried out in February 2026 among 174 licensed mental health clinicians based in the United States, found that 61.07% believe the use of AI emotional-support tools often (18.12%) or sometimes (42.95%) contributes to patients postponing licensed mental health treatment.
These findings arrive amid ongoing concerns about limited access to mental health care. The 2024 National Survey on Drug Use and Health reported that nearly half (47.9%) of American adults experiencing mental illness did not receive any form of mental health treatment within the past year.
Clinicians participating in the survey expressed notable concern about the role of AI tools in emotional support. On average, respondents rated their concern at 3.58 out of 4. Furthermore, 44.83% indicated they were aware that some of their clients are using AI-powered tools such as chatbots, mobile applications, or virtual assistants to seek emotional or mental health guidance before therapy or alongside it.
Patient transparency about AI use appears inconsistent. In the previous 12 months, 9.83% of clinicians reported that patients regularly disclose using AI tools for emotional support. Another 28.90% said such disclosures occur occasionally, while 20.81% stated that patients rarely share this information. Meanwhile, 40.46% reported no disclosure of AI use at all.
AI-based support tools appear most commonly used among younger adults. Among clinicians who were aware of AI usage, 55.46% reported patients aged 26 to 40 using such tools. Patients aged 18 to 25 followed at 42.86%. Usage was also identified among patients aged 41 to 60 (32.77%), under 18 (19.33%), and those aged 61 and above (5.88%).
Respondents said continuous availability was the primary reason patients use AI tools, with 64.24% identifying round-the-clock access as the main factor. Additional reasons included lower costs (37.75%), reduced intimidation compared with speaking to a clinician (31.13%), easier access than booking therapy (30.46%), faster responses (29.14%), anonymity (27.81%), difficulty locating providers (25.17%), insurance barriers (15.89%), and negative past therapy experiences (9.27%).
Emily Mendenhall, Professor and Medical Anthropologist at Georgetown University, said the findings reflect broader structural challenges within the U.S. mental health system.
“Mental health care in the United States is only getting more difficult to access,” she said. “Because of structural barriers and rapid shifts in AI, the low-cost immediate strategy of AI as therapist may seem like a replacement for people who are struggling and cannot access the care they need.”
Dr. October Boyles, DNP, MSN, BSN, RN, behavioural health expert and clinical consultant at ICANotes, noted that AI tools should not replace licensed clinical care.
“When individuals delay seeking professional care, especially for moderate to severe symptoms, opportunities for early intervention can be missed,” said Dr. Boyles. “Technology can support clinicians and patients, but it must be implemented thoughtfully, with patient safety and evidence-based practice at the forefront.”
The survey gathered responses from 174 licensed U.S. mental health clinicians and was conducted during February 2026.