Luke Ralston, a biomedical engineer and scientific reviewer at the FDA, emphasized the importance of transparency in AI regulation during a presentation at the HRX conference in Atlanta. The FDA sees AI regulation as specific to each device and its intended use. Two major issues in healthcare AI are performance drift and data generalization. Companies should conduct post-market monitoring to track model performance in real-world settings and gather diverse, representative data for training. Ralston stressed the need for large, clean datasets, highlighting the importance of considering various factors beyond demographics. As AI evolves, healthcare companies must address these challenges to ensure safe and ethical deployment.
Source link
These Are 2 of The Biggest Issues with Healthcare AI, According to an FDA Reviewer
