Serial entrepreneur Munjal Shah has launched a new startup called Hippocratic AI that aims to leverage the power of large language models (LLMs) to improve patient outcomes in areas like chronic care management and patient navigation.
Shah sees an opportunity for generative AI to help fill healthcare gaps caused by shortages of nurses and other medical staff. There are an estimated 100,000 fewer nurses today than a couple of years ago, with over 600,000 more expected to leave the field by 2027. Meanwhile, aging populations are driving increased demand for care. This is where Munjal Shah believes AI can make a real difference.
As the founder of companies acquired by Google and Alibaba, Shah has experience applying machine learning to e-commerce. However, the recent advances in generative AI enabled by large language models are even more promising for the healthcare industry. Unlike the classifier AI used in e-commerce, LLMs can generate personalized responses tailored to individual patients’ needs.
Shah wants to use LLMs to “10x the number of healthcare workers” through what he calls “super staffing.” The goal is not to replace human nurses but to multiply their impact by automating routine communications. This could include handling tasks like sending chronic care reminders, providing diet advice, booking specialist appointments, and explaining insurance statements. Munjal Shah believes offloading these responsibilities to AI will reduce nurse burnout while expanding access to care.
Crucially, Hippocratic AI will avoid diagnostic applications due to continued concerns around AI hallucination. Shah acknowledges it would be “unsafe” to rely on generative AI for patient diagnoses at this stage. However, there are still ample opportunities to deploy LLMs in non-clinical roles – as virtual assistants rather than robot doctors. Before launching any new AI-powered services, they will be rigorously evaluated by human medical experts.
By leveraging machine learning to handle more routine administrative and educational responsibilities, Shah aims to create capacity for human healthcare professionals to focus on higher-value tasks. This hybrid model combines the communicative abilities of AI with specialized human skills and ethical judgment. The end goal is not to replace nurses but to provide the level of personalized attention and care for every patient that would be possible with drastically expanded staffing.
Early pilot studies have shown promising results, with LLMs demonstrating the ability to comprehend complex medical literature and tailor explanations to individual contexts. However, Shah emphasizes human oversight is still essential for now. Every new application will require extensive real-world validation before deployment.
If generative AI can successfully automate routine communications while maintaining accuracy, Hippocratic AI envisions a future where every patient has 24/7 access to a personalized virtual health assistant. This AI support structure would improve health outcomes and provide companionship for isolated groups like older people.
However, many open questions remain about data privacy, algorithmic bias, and how to balance automation with human empathy. Shah believes the only way to integrate AI into healthcare responsibly is by proceeding cautiously and collaboratively every step of the way.
By focusing first on supplementary use cases instead of mission-critical diagnoses, Hippocratic AI hopes to build trust through iterative improvement. While the technology still warrants some healthy skepticism today, Munjal Shah believes LLMs represent the most promising path yet toward democratizing access to quality healthcare globally. If generative AI can help fill the widening gaps in the healthcare workforce, we will have witnessed both a technological and a humanitarian breakthrough.