- Hippocratic AI’s product will focus on non-diagnostic apps to improve access to healthcare
- The company's language model outperformed GPT-4 on 105 of 114 healthcare certifications and tests
Hippocratic AI has announced the industry’s flagship safety- and healthcare-focused Large Language Model (LLM) and a $50 million seed funding round co-led by Andreessen Horowitz (a16z) and General Catalyst, Bankless Times learned from a press release.
Improving patient access and safety
Researchers have shown that LLMs like ChatGPT and GPT-4 can pass the USMLE (US Medical Licensing Exam), but no company had developed a program specifically for healthcare applications before. Hippocratic AI’s product will focus on non-diagnostic apps to improve access to healthcare and ensure patient safety.
Hippocratic AI was founded by a group of AI researchers, doctors, medical administrators, and Medicare professionals from leading international institutions. Julie Yoo, General Partner at Andreessen Horowitz said:
We believe Hippocratic AI’s cross-disciplinary, safety-first approach is what the healthcare industry needs to be able to maintain trust in the power of responsible deployment of generative AI solutions.
Hemant Taneja, CEO and Managing Director at General Catalyst, commented:
The healthcare industry needs its own AI platform, one that is focused on empowering the workforce, reducing burnout, and improving patient safety and experiences. We joined forces with the Hippocratic AI team, our health assurance ecosystem, and the a16z team to build this platform.
Hippocratic outperformed GPT-4 on tests
To build a safer large language model, Hippocratic AI focused on certification and human feedback, among other things. The company tested its language model on 114 healthcare certifications and tests and outperformed GPT-4 on 105 of them.
Molding AI via human feedback
LLMs have an approach to “molding” AI via feedback from humans called Reinforcement Learning with Human Feedback (RLHF). According to majority opinion, this approach was behind ChatGPT’s superiority to previous versions of creator OpenAI’s language models.
Hippocratic AI worked with healthcare professionals to help educate and guide the LLM. They did this by giving ratings to the program’s responses.
