With Great Power, Comes Great Responsibility: Unpacking Views on Using AI for Mental Health Care 

A new study led by Columbia University School of Nursing examines patient perspectives on using artificial intelligence (AI) in mental health care. The survey, involving 500 U.S. adults, found that 49.3% of participants viewed AI as beneficial for mental health care. African Americans and individuals with lower self-rated health literacy were more likely to hold this view, while women were less likely. Additionally, 81.6% of respondents believed health professionals should be held accountable for AI-related misdiagnoses in mental health. The study, titled “Patient Perspectives on AI for Mental Health Care: A Cross-sectional Survey Study,” was published in JMIR Mental Health.

The study uncovered key patient concerns, including fears about incorrect diagnoses, inappropriate treatments, reduced interaction with care providers, and potential breaches of confidentiality. Participants also expressed a desire for greater transparency, including visibility into how AI tools are applied, understanding their performance, and maintaining autonomy in care decisions. “Understanding patient perceptions of if and how AI may be appropriately used for mental health care is critical,” said Natalie Benda, PhD, an assistant professor at Columbia Nursing who led the study. “Our findings can support health professionals in deploying AI tools safely.”

Based on the findings, the researchers recommend that health professionals rigorously test AI tools in clinical simulations before broader use, ensure transparency about AI’s accuracy and risks, address biases, and communicate clearly about AI’s role in care. They also stress the importance of fostering patient autonomy and collaborative decision-making.