Inside AI in Patient Care: Columbia Nursing Study Reveals What Patients, Providers, and Developers Really Think
A new Columbia University School of Nursing study examines perspectives from patients, providers, and developers on how artificial intelligence (AI) is implemented in health care. Through 34 interviews, the researchers found that all groups emphasized the need for AI to provide actionable benefits while minimizing bias, stigma, anxiety, and harm.
Particularly, patients expressed interest in professional interpretation of AI outputs due to perceptions of how AI could change care trajectories and lead to unexpected interventions when dealing with sensitive information and conditions such as mental health. For example, a patient expressed concerns that AI’s mental health predictions and interpretations could lead to child protective services intervention:
“I may run a 100 score and it’s like, ‘oh i passed’. Like, why is the doctor coming to talk to me? And ACS (local CPS department) is coming to take my child?"
Another key concern centers on the training data used for AI models and how incomplete or outdated patient information can undermine the accuracy of predictions. Health professionals shared how certain data, or the lack thereof, could lead to biased or uncomprehensive results. Developers felt responsible for making systems with high predictive performance that also mitigated potential biases.
“Our findings suggest that AI implementation in health systems may require coordination between patients, health professionals, developers, and health systems administrators and leaders… Solutions exist through established educational pathways and participatory design methods. However, they require resource allocation and cultural shifts in AI development practices,” the authors conclude.
The study, “A Qualitative Interview Study Investigating Patient, Health Professional, and Developer Perspectives on Real-World Implementation of Patient-Centered AI Systems,” was published in npj Digital Medicine on May 5, 2026.