Hippocratic AI, a company specializing in safety-focused large language models (LLMs) for healthcare, has been granted its first U.S. patent for its Polaris architecture. Polaris combines a primary LLM, trained for clinical safety and patient compliance, with support models that assess patient responses to prevent harmful actions, such as incorrect medication intake. The system is designed for various healthcare applications, including medication onboarding, EHR assistance, and chronic care management. CEO Munjal Shah highlighted the significance of the patent, emphasizing the innovative “constellation architecture” of multiple reinforcing LLMs that ensure the safety and reliability of the system. Since its launch, the Polaris architecture has influenced other Agentic AI systems in the healthcare domain.
Hippocratic AI has been expanding its influence through partnerships and funding. It collaborated with Adtalem Global Education, Chamberlain University, and Walden University to create curricula and certifications for training clinicians to use AI effectively. Additionally, the company established a nurse advisory council to complement its physician advisory council, focusing on equity, workforce capacity, and improving healthcare outcomes. Since its founding in May 2023, the company has raised $137 million in funding, including a $53 million Series A round and additional investments from NVIDIA’s NVentures and others, elevating its valuation to $500 million. These initiatives position Hippocratic AI as a leader in integrating AI into healthcare while prioritizing safety and accessibility.