Eric Wu

AI development in healthcare has made impressive technical strides in recent years — however, a significant challenge that remains is ensuring its safe and effective deployment in healthcare settings. It is well-known that AI models often fail to generalize to patient populations that they have not previously seen. Additionally, the pace of regulation and evaluation standards for medical AI have lagged behind the breakneck pace of technological advancements. My research focuses on a) advancing a data-driven understanding of the regulatory landscape for medical AI in the US and b) developing computational methods for assessing and reducing biases in the evaluation process.