At the recent 2023 HLTH conference, Munjal Shah, CEO of Hippocratic AI, spoke about how artificial intelligence could help alleviate worldwide healthcare staffing shortages through “understaffing.”
The annual HLTH event in Las Vegas brings together leaders in healthcare innovation. This year’s key topic was leveraging generative AI to improve patient care and outcomes. During the “There’s No ‘AI’ in Team” panel, Shah argued that while diagnostic applications of AI remain risky, tremendous potential exists for AI as virtual “staff” in non-clinical roles.
The panel’s premise was that an optimal approach combines human expertise with AI capabilities. Shah believes generative language models trained on medical conversations could provide services like chronic care nursing, appointment coordination, and test result explanations at a fraction of the cost of human staff. This AI “understaffing” could dramatically expand healthcare access and equity.
The World Health Organization projects a shortfall of 10 million health workers globally by 2030. Shah insists this is already impacting healthcare systems and underserved patients. He envisions AI filling gaps in the overstretched workforce, working alongside human providers in a “centaur” model.
Shah says deep collaboration with medical institutions is crucial to building safe, trustworthy AI for healthcare. Thousands of clinicians are training and testing Hippocratic AI’s systems. The key is reinforcing expert, evidence-based responses through reinforcement learning and human feedback.
While diagnostic AI remains risky, Shah sees significant potential in using conversational AI for patient interactions. At just $1 per hour, AI “staff” could provide services like post-discharge follow-ups at a scale impossible for human workers. The goal isn’t replacing people but augmenting human capabilities.
Some applications Shah highlighted include explaining billing, providing genetic counseling, answering surgery questions, and delivering test results. A recent JAMA study even found ChatGPT responses were preferred to doctors’ on measures of quality and empathy.
The low cost of AI conversational agents means providing dedicated support for all patients with multiple chronic conditions, something unrealistic for human staff. Shah believes this “understaffing” can help cover unmet needs due to overstretched and under-resourced care systems.
In Shah’s view, generative AI’s strengths in versatile conversation and reasoning across documents make it ideal for patient-facing interactions. Training AI in high-quality medical conversations allows for replicating the human touch unlimitedly.
Shah acknowledges AI isn’t a panacea for healthcare’s challenges. But thoughtfully applied, he believes it can make fundamental differences for underserved populations by expanding access to support services requiring personalized interaction.
Other panelists agreed that human-free AI solutions are unlikely to fail in medicine. Partnerships maximizing the complementary strengths of both human expertise and AI capabilities could lead to better patient experiences, equity, and outcomes.