In this talk, I draw on Feenberg’s philosophy of technology to assess the increasing challenges to governing the personal health information essential for advancing AI machine learning innovations in healthcare. I highlight the opacity, scale, and scope of PHI data used by deep learning algorithms as novel challenges to health data governance, discussing emerging uses such as personally generated health data (PGHD) and precision genomics. I argue that these characteristics of machine learning are already beginning to overwhelm existing data governance approaches such as privacy and security regulations and informed consent. Using Feenberg’s philosophy of technology, I argue that technical codes of current and emerging AI health systems suppress other potentialities, and I explore other sociotechnical alternatives.