Serial entrepreneur Munjal Shah sees vast potential in using large language models (LLMs) to provide supplemental health care services, though not for diagnosis. With his latest venture, Hippocratic AI, Shah aims to harness conversational AI to help address the deepening global shortage of health workers.
Shah has an extensive background in AI, noting he built his first 16-node neural network in 1992. Having tracked the evolution of the technology, he believes recent leaps in generative AI represent a “true breakthrough” that’s actually “underhyped.” Unlike past AIs focused on classification and categorization, modern language models can generate new content while capturing human communication patterns.
According to Shah, many saw conversational tools like ChatGPT and assumed AI could replace doctors. Shah staunchly disagrees, saying blind trust in an LLM’s capabilities invites potentially deadly mistakes. However, he realized there’s far more to healthcare than diagnosis, identifying a significant opportunity in supplemental care.
With over 15 million unfilled health worker positions globally, providers struggle to meet demands in nursing, dietetics, navigation, and other critical services. That’s where Hippocratic AI comes in, leveraging LLMs to provide personalized support efficiently. As Shah explains, an LLM could regularly check if patients are taking medications, making appointments, securing rides, or getting enough food – asking questions and assisting where needed.
Crucially, Hippocratic AI converses in a knowledgeable yet empathetic tone modeled after human interactions. Shah stresses that patients need to want to engage with the LLM for it to be helpful. This requires extensive feedback from medical professionals during training to capture the right communication style.
The goal is “super staffing” – using AI to assign a dedicated assistant to every patient affordably. With unlimited time and no risk of burnout, LLMs could provide consistent chronic care management that most practices cannot match today. Shah envisions these AI agents eventually helping fill health worker shortages worldwide.
While acknowledging risks in overestimating LLMs’ abilities, Shah believes narrowly targeted applications in supplemental care could provide tremendous value. With thoughtful implementation, conversational AI might soon play a supporting role in helping patients worldwide receive the ongoing assistance they need.