At this year’s NVIDIA GTC conference, Munjal Shah, founder of Hippocratic AI, unveiled one of the pioneering real-world deployments of large language models (LLMs) in the healthcare domain. Hippocratic AI has spent over a year meticulously crafting “empathetic” artificial intelligence agents capable of engaging in nuanced voice interactions with patients.
The overarching goal is to safely implement these conversational AI assistants to perform various non-diagnostic tasks, thereby alleviating the immense strain on human clinicians caused by staffing shortages. With over $120 million in backing from prominent investors like General Catalyst, Andreessen Horowitz’s Bio + Health fund, and Premji Invest, Hippocratic AI’s solution aims to offload routine responsibilities to its AI agents. These agents can provide pre and post-operative guidance, gently encourage patient adherence to care plans, and respond to medication-related queries while maintaining a friendly, personalized, and caring demeanor.
As Munjal Shah articulated in a recent statement regarding AI’s partnership with NVIDIA to reduce response latency, “With generative AI, patient interactions can be seamless, personalized, and conversational. In order to have the desired impact, the speed of inference has to be incredibly fast.”
NVIDIA’s powerful AI accelerators can help facilitate these conversations at a natural, human-like cadence. According to Shah, every half-second reduction in latency can boost patients’ perceived emotional connection by up to 10%. He underscored that “NVIDIA’s technology stack is critical to achieving this speed and fluidity,” referring to Hippocratic AI’s “empathy inference engine” that enables low-latency AI interactions.
In an interview with UCSF’s Rosenman Institute, Shah contrasted the current state of conversational AI with outdated interactive voice response (IVR) systems, stating, “The problem with older IVR systems is that the comprehension is very low. If you don’t say it in exactly the right way, it doesn’t work at all, right?” He highlighted the remarkable progress, asserting, “You have to think about the old chatbot was IQ 60, [and] this one is IQ 130. It’s a very different level of comprehension.”
Hippocratic AI’s service endeavors to mitigate the crippling staffing shortages plaguing the healthcare system by enabling human nurses and doctors to delegate routine tasks to generative AI agents. The company’s mission is to increase access to healthcare services and empower nurses to allocate more time to higher-risk clinical care and personal interactions that AI cannot replicate.
This pivotal advancement was made possible by recent breakthroughs in large language models like OpenAI’s GPT and Anthropic’s Claude, versatile AI systems that train text corpora to engage in open-ended dialogue. The Hippocratic approach emphasizes serrations, adhering to the principle of “no harm.” This entails creating LLMs exclusively trained on authoritative, evidence-based medical sources and subjecting them to rigorous reinforcement learning overseen by human medical professionals.
HippocratiAI’s agents are designed solely for straightforward, non-diagnostic workflows that reduce routine follow-up orrative burdens, not for high-stakes medical advice. Initial use cases include pre and post-operative patient guidance, onboarding for new medications, and adherence reminders.
With the U.S. Bureau of Labor Statistics projecting a need for 275,000 additional nurses by 2030 to meet the intensifying care demands of an aging population, Hippocratic AI’s “empathetic agents” present a promising solution. By automating time-consuming tasks, these AI assistants could potentially alleviate burnout among overworked nurses while elevating the human aspects of their roles, such as urgent care, emotional support, counseling, and addressing complex diagnostic and treatment inquiries.
