This episode’s guest Dr. Enes Hosgor is driven by his experiences in regulated industries and a desire to move beyond "toy algorithms" in medical AI. He stresses the importance of investing in infrastructure to address the challenges of generalizability and reproducibility in real-world applications. In his article on bias in medical AI, Enes highlights the need for novel regulatory approaches that allow stakeholders, including physicians and healthcare systems, to understand and trust AI predictions without requiring extensive technical expertise. Enes also discusses the challenges of defining regulatory boundaries around AI models and the iterative nature of incorporating relevant data.