Serial entrepreneur Munjal Shah sees vast potential in using large language models (LLMs) for non-diagnostic health care services. While risky for diagnosis, he believes conversational AI could help address the massive global shortage of health workers.
Munjal Shah noted that recent leaps in generative AI lead to engaging chatbots. Still, applying technology in health care requires care to avoid potential harm from AI limitations. LLMs can accumulate knowledge rapidly and discuss intelligently but sometimes present false information as accurate.
So rather than diagnosis, Shah targeted chronic care support applications at his new startup, Hippocratic AI. The LLMs conduct helpful, non-clinical dialogs on diet, appointments, transportation, and more. The goal is to reduce burdens on human specialists through responsive, cost-effective AI assistance.
The founder called generative AI “a true breakthrough,” saying it is even “underhyped” given its original, human-like content creation. Unlike past “classifier” AI for sorting fixed data, today’s models can learn structures to generate new materials. Hippocratic AI’s system is expressly trained in medical language to ensure empathetic patient conversations.
Shah believes this innovation enables “super staffing” that expands access despite limited workforces. While ideal care includes extensive discussions, professionals cannot devote endless hours to each person. AI has that capacity, delivering personal chronic support at scale.
So Hippocratic AI wants to create “fully autonomous agents” handling health services worldwide. Even essential assistance like medication reminders can profoundly impact outcomes. And the tech’s conversational skills suit those applications better than strict triage or diagnoses.
Ultimately, Shah sees vast possibilities if care is taken around regulation. Health LLMs cannot replace human judgments but may expand their reach. By handling routine information flows, AI could let professionals focus their talents where uniquely human understanding is essential. Then, augmented staffing can meet more needs and improve more lives.
