At the Future of Health 2023 conference, Hippocratic AI founder Munjal Shah laid out an ambitious vision for how artificial intelligence could help resolve an escalating crisis facing healthcare systems around the world: staffing shortages. He explained how his company’s large language model (LLM), trained on medical data and rigorously evaluated by health professionals, is poised to enable a new super-staffing model that provides patients with their own personal 24/7 health assistant.

The Gathering Storm of a Healthcare Staffing Crisis

Munjal Shah took the stage at Future of Health 2023 against a backdrop of a gathering storm in healthcare. Staffing shortages resulting from an aging population with heightened care needs, compounded by burnout and loss of medical professionals, have reached crisis proportions across healthcare systems globally.

In the U.S. alone, nearly half a million nurse vacancies are projected by 2031. Up to 124,000 doctor shortfalls could materialize in the next decade. Worldwide, the World Health Organization forecasts a gap of 10 million health workers by the end of this decade. Without intervention, these deficits threaten to degrade quality of and access to care at a time when both are more vital than ever.

For Shah, this crisis demands fresh thinking and new solutions. “We need [increase by] 10 times the number of health care workers,” he told the audience. “You can only do that with autopilot.” The vision Shah laid out is one enabled by the unique capabilities of modern artificial intelligence — specifically, large language models.

The Promise and Limitations of Healthcare AI

Shah explained a taxonomy for health AI, dividing applications into three categories: co-piloting, autopiloting, and super-staffing. Co-pilot systems keep health professionals involved by offering assistance like report drafting and treatment research. While beneficial, co-pilots’ requirement of human oversight limits potential efficiency gains. Fully autonomous systems without humans in the loop comprise autopiloting applications. Appropriate only in low-risk contexts, their hands-off operation unlocks vastly greater leverage.

Super-staffing represents the logical conclusion of autonomous AI in healthcare: mass deployment of low-cost virtual health assistants providing round-the-clock support directly to patients. For Munjal Shah, super-staffing via AI is the most viable path to the Millennial-era vision of universal 24/7 healthcare. Modern large language models finally offer the cost structure, versatility, and performance needed to make this vision a reality.

Development Driven by Healthcare Professionals

Maintaining healthcare professionals’ trust in and oversight of AI is central to Hippocratic AI’s development process under founder Munjal Shah. While supremely capable, LLMs intrinsically lack capacities like self-reflection that help humans handle high-risk situations responsibly. Hence why diagnostic applications remain off limits.

However, constantly improving performance on low-risk tasks makes LLMs ideal for the high-value nursing, administrative, and consumer-facing jobs contributing to care access issues. Hippocratic AI’s LLM undergoes rigorous medical certification testing, recently outscoring GPT-4 on most evaluations. More importantly, the model is trained on real clinical data then refined based on regular structured feedback from thousands of doctors, nurses, and other stakeholders. Shah stated that each LLM role launches only once these professionals confirm its readiness.

Democratizing Healthcare Through AI Super-Staffing

The limitations constraining human-based healthcare motivates the super-staffing approach Hippocratic AI is pioneering under founder Munjal Shah. Humans can only see so many patients; systems don’t tire. Humans need to sleep; AI works 24/7. Humans have language barriers; LLMs speak every tongue. And crucially, humans costs orders of magnitude more than software to employ.

Shah cast the dichotomy in terms that every patient can understand: “Why don’t we have 350 million health care workers for 350 million people? It would lead to better outcomes.” Of course, America can’t feasibly train that many new doctors and nurses. But AI makes a heretofore impossible 1:1 care ratio suddenly within reach. Integrating autonomous systems for high-value tasks promises to boost staff productivity markedly according to Shah. However, the real magic lies in allowing each person their own dedicated health assistant.

This personal assistant handles the many small but vital jobs that overworked staff currently deprioritize. Lingering questions get answered. Appointment reminders assure attendance. Visit summaries with follow-up actions better adherence. Ongoing nursing ensures proper recovery and health maintenance. AI makes this degree of high-touch management feasible by obviating the need to schedule in-person visits for each concern.

Munjal Shah’s expansive vision leverages AI’s unique strengths to provide universal 24/7 personalized care at scale — democratizing healthcare by effectively putting a nurse in every patient’s pocket. Of course, responsible oversight and evaluation remain imperative every step of implementation under Hippocratic AI’s methodology. But by collaborating closely with health professionals, Shah aims to transmute AI’s abundant challenges into abundant help for all.