Reducing Racial Biases within Healthcare Applications of Artificial Intelligence (AI) With Transparency

Artificial intelligence (AI) is increasingly being used in healthcare for applications such as drug discovery, diagnostics, disease management, and delivery of services. However, integrating AI and healthcare raises concerns about reinforcing existing societal prejudices: AI systems are known to exhibit racial biases by making inaccurate and unreliable decisions based on race when it is irrelevant to the task. Furthermore, government directives currently lack consistent standards for regulating AI and offer insufficient guidance on preventing the perpetuation of harmful racial biases, especially in healthcare. To improve patients’ quality of life interacting with AI systems, it is essential to ensure transparency regarding these systems. Additionally, it is vital to ensure that innovation dedicated to improving healthcare enhances the integrity of the patient’s experience rather than compounds existing systemic disparities. The authors propose three recommendations to address racial biases in healthcare applications of AI and emphasize the need for legislation placing AI regulation in healthcare at the forefront of healthcare policy agendas.