By the Numbers: 27 April 2026
22.2%
The percentage of 18- to 21-year-olds.in the U.S. who have used "generative artificial intelligence (AI) as a mental health resource," according to an article in JAMA Psychiatry -- Patients Use AI—Clinicians Should Ask How.
US individuals are turning to generative artificial intelligence (AI) as a mental health resource. More than 5 million US youth (13.1%) have sought mental health advice from AI, with rates reaching 22.2% among 18- to 21-year-olds. A 2025 study of adults with mental health conditions who use large language models reported nearly half use them for support, including for anxiety, depression, and personal advice. Users seek emotional support, companionship, psychoeducation, and help processing difficult experiences, likely between sessions and sometimes instead of clinical care altogether. While professional guidance focuses on how clinicians should use AI tools, conversations with patients about their AI use receive less attention. We argue that such conversations are essential and that clinicians already have the skills these conversations require.
22.2%
The percentage of 18- to 21-year-olds.in the U.S. who have used "generative artificial intelligence (AI) as a mental health resource," according to an article in JAMA Psychiatry -- Patients Use AI—Clinicians Should Ask How.
US individuals are turning to generative artificial intelligence (AI) as a mental health resource. More than 5 million US youth (13.1%) have sought mental health advice from AI, with rates reaching 22.2% among 18- to 21-year-olds. A 2025 study of adults with mental health conditions who use large language models reported nearly half use them for support, including for anxiety, depression, and personal advice. Users seek emotional support, companionship, psychoeducation, and help processing difficult experiences, likely between sessions and sometimes instead of clinical care altogether. While professional guidance focuses on how clinicians should use AI tools, conversations with patients about their AI use receive less attention. We argue that such conversations are essential and that clinicians already have the skills these conversations require.