Safety-focused large language model company Hippocratic AI has been issued its first patent by the U.S. Patent Office, which includes the company's LLM innovations incorporated into its safety-focused LLM built with a constellation architecture dubbed Polaris.

Hippocratic AI's generative AI solution focuses on nondiagnostic healthcare tasks such as medication onboarding and monthly reconciliation, hospital and payor policy questions, and EHR assistance. 

The patent covers the company's Polaris system, which includes its primary model and support models that make up its low-latency conversational AI system tailored for healthcare applications.

The company's primary model includes its conversation interface, a custom-trained LLM that uses clinical techniques to ensure patient compliance. 

Its support models include various LLMs that analyze patient responses to ensure safety and that the patient has not indicated they will knowingly or unknowingly perform harmful activities to themselves, such as exceeding the medication intake amount recommended by their provider. 

Specific applications in the patent include intake tasks, preventative screenings, scheduling and waitlisting tasks, preoperative tasks, discharge tasks, billing insurance and chronic care-management tasks. 

"Our first granted patent! ... And it is an important one. At the heart of the safety of our system is the constellation architecture. Our system of multiple reinforcing LLMs is how we got our product to be extremely safe," Munjal Shah, cofounder and CEO of Hippocratic AI, said in a LinkedIn post.  

"Since we launched our Polaris 1.0 (and now 2.0) architect, it has become a technique used by many other Agentic AI products. We are proud to say the USPTO recognized our innovation and pioneering status. I wanted to commend our crazy smart and creative research team on this patent." 

THE LARGER TREND

In September, Hippocratic AI announced through MobiHealthNews that it was partnering with Adtalem Global Education to create an academic curriculum to train and certify clinicians on the use of AI in healthcare.

The company partnered with Adtalem's Chamberlain University and Walden University to jointly develop a curriculum and certification program that helps nurses understand how to use, integrate and oversee AI systems effectively in healthcare. 

In February, the company also announced it was launching a nurse advisory council that will work in conjunction with its physician advisory council to ensure the safe development of its LLMs aimed at increasing care equity, the supply and availability of healthcare workers, and better healthcare access and outcomes.

The company launched in May last year with $50 million in seed funding in a round coled by General Catalyst and Andreessen Horowitz. The company added $15 million in funding to its seed round two months later. 

Earlier this year, Hippocratic AI closed an oversubscribed $53 million Series A funding round, bringing its valuation to $500 million.

Last month, the company announced it added $17 million in investment to its closed extended Series A round with participation from NVentures (NVIDIA's venture capital arm), Greycroft and LeoShapiro of 7Wire Ventures. The additional investment brought the company's total raise to $137 million.