Munjal Shah brings over 25 years of AI expertise to his new startup, which targets supplemental healthcare services. He sees today’s large language models as a breakthrough in realistic human conversation – but cautions against unleashing them directly on diagnosis. Shah’s approach with Hippocratic AI focuses first on expanding chronic care support to more patients through non-diagnostic roles.
The global health worker shortage has reached crisis levels, estimated at 15 million unfilled roles. Patients, especially those with chronic conditions, can’t access the guidance they need on issues like medications, appointments, transportation, and nutrition. Generative AI offers a path to provide personalized assistance cheaply and at an infinite scale.
Shah distinguishes classifier AI, which categorizes data for diagnostic decisions, from creative models like ChatGPT that generate new content. The latter better suits open-ended patient conversations. Proper training remains vital so patients enjoy the exchange – Shah points out ChatGPT still “talks like it’s writing a blog.”
Hippocratic AI conversations collect no patient data. The system helps guide behaviors, freeing human care workers up for higher-level duties. Shah sees an opportunity to vastly expand access to supportive interactions without diagnosing.
This “super staffing” may one day deliver health guidance to patients worldwide. For now, Hippocratic AI focuses on supplemental chronic care conversations, helping patients follow care plans while avoiding the risks of AI diagnosis and bridging significant gaps in the care journey even without replacing human roles.
Shah brings the technical background to apply generative AI’s creative potential responsibly. While the hype and risks rightly focus on diagnosis, huge unmet needs exist in ongoing care. With Hippocratic AI’s thoughtful approach, AI conversation could give patients the next best thing to having a full-time nurse.