Munjal Shah, co-founder and CEO of Hippocratic AI, explains how the company handles safety in clinical AI. Hippocratic deliberately keeps its scope to what a registered nurse can do, so it avoids diagnosis and prescribing. It also uses a “scope safety” approach in which 31 AI models work together to supervise the main model, rather than relying on one large model. Shah says this setup catches errors, such as an unsafe medication suggestion for a patient with chronic kidney disease. He also takes a firm position on accountability, saying “you cannot build a clinical AI without taking liability.” Hippocratic’s contracts let health systems pursue legal action against the company, and it maintains what he calls a “fortress balance sheet” to cover damages beyond insurance limits.
On the workforce, Shah argues AI will expand healthcare employment rather than shrink it. America’s aging population needs roughly ten times more care than younger adults, and nursing shortages already exist. He says Hippocratic’s most successful use cases are new services that were previously too costly to offer. One example is calling 50,000 at-risk patients during a heat wave to give them preventive guidance.