Why human review is key to the success of AI in health care
Why human review is key to the success of AI in health care
Publish Date: 2026-05-08 12:01:00
Source Domain: health.ucdavis.edu
-
Increasing AI Use in Healthcare: AI tools in healthcare are being increasingly utilized to analyze medical images, predict risks, and monitor patient conditions remotely. However, biases can occur due to unrepresentative or unbalanced data.
-
Human Oversight for Bias Reduction: The study led by UC Davis Professor Courtney Lyles underscores the importance of human oversight to minimize AI bias, ensuring the safety and reliability of AI decisions.
-
Interdisciplinary Review Panels: This study recommends the use of interdisciplinary review panels that combine medical, epidemiological, behavioral science, engineering, and data science expertise to assess AI decisions and bias in explainable AI models.
-
Explainable AI Importance: Explainable AI (XAI) involves understanding the underlying reasons for AI model predictions, which is crucial for validating AI decisions and highlighting potential biases.
-
Importance of Diverse Input: Including community members and patient advocates in the review process adds valuable, lived experience insights that can help ensure that AI tools better reflect the communities they serve.
-
Creating Teams for Ethical AI: The study suggests the formation of dedicated teams with diverse expertise to oversee AI deployment in healthcare, ensuring accurate and contextual results.
-
Private-Public Partnerships: To advance equitable AI development, private-public partnerships are necessary, exemplified by initiatives like UC S.O.L.V.E Health Tech, that integrate research with practical implementation by industry partners.
-
Implementation at UC Davis Health: UC Davis Health has implemented various AI initiatives, including AI governance committees, equitable rollout processes, AI Scribe for clinical documentation, and efforts to reduce bias in AI predictive models.
