A new episode of our Twin Things Podcast is now available to watch. In episode #8, “Can AI Really Think Like a Doctor?”, host Kasia Baliga-Nicholson sits down with Dominik Stosik, Clinical Advisor for AI & Digital Health Solutions at Sano, to ask a deceptively simple question: what if the biggest challenge for AI in medicine is not finding the right answer, but knowing how doctors actually think?<\/p>\n\n\n\n
Their conversation moves through large language models, clinical reasoning, radiology and the real limits of using AI in medical settings. From emergency rooms and medical imaging to hallucinations, deepfakes, cognitive bias and hospital workflows, they explore why AI can be shockingly good at pattern recognition yet still struggle with the messy, uncertain and deeply human process of diagnosis. The takeaway is a compelling one: AI should not simply replace doctors’ thinking, but become a kind of “mental exoskeleton” that helps clinicians question, verify and make better decisions.<\/p>\n\n\n\n