There’s a newcomer artificial intelligence company on the block that wants to introduce the power of a conversational chatbot for the healthcare industry: Hippocratic AI, which just launched with $50 million in funding co-led by General Catalyst and Andreessen Horowitz.

Founded by a group of physicians, hospital administrators, Medicare professionals and medical researchers, Hippocratic is named after the oath that doctors take to “do no harm.” The artificial intelligence model that the company is developing is designed to be the first safety-focused large language model specifically focused on healthcare applications.

Although there have been multiple AI models on the market that have shown great promise in healthcare capacities, none of them has created a commercial model specifically tuned for healthcare applications. Although researchers have already shown great strides in creating models that show promise in healthcare, such as Google LLC’s Med-PaLM 2, which passed the U.S. Medical Licensing Exam.

Of course, simply passing exams isn’t enough to show that an AI is safe enough to be involved in healthcare. The regulatory landscape involving artificial intelligence and LLMs is also rapidly evolving as more AI models are being produced and integrated into even more industries. Healthcare in particular is very sensitive to patient privacy and safety, including regulations such as the Health Insurance Portability and Accountability Act.

To start, Hippocratic’s AI model will not be involved in diagnosing medical conditions or directly with patients. That will allow the company to work on building the necessary safety features and guardrails for the chatbot. As has been seen, generative AI models such as ChatGPT can be prone to errors, known as “hallucinations,” where the AI will confidently make false statements.

In order to reduce the chances of its model producing false information, co-founder and Chief Executive Munjal Shah said, the company engages with healthcare professionals to help guide and train the artificial intelligence model. The technique to do this is called reinforcement learning with human feedback. It uses a pool of experts in the field to review responses from the AI model who rank replies as good or bad that align its future replies to reduce the chances of errors.

“We aren’t just saying these professions will help us evaluate our system,” said Shah. “We are saying we won’t launch each unique role for the LLM unless the professionals who do that exact task today agree the system is ready and safe.”

Some of the roles that the company’s artificial intelligence assistant could fill in the future could include a patient navigator, dietician, genetic counselor, enrollment specialist and medication reminder.

The company is also focused on making its AI model more humanlike in its conversational ability, what Hippocratic calls its “bedside manner,” because it will be working with people in the healthcare industry. In the same way most doctors and nurses approach patients with empathy and compassion, the AI should present itself with an air of not just confidence but care and concern that sets patients at ease.

“In healthcare settings, it isn’t just important to answer the patient accurately. It is equally important that it is done with great bedside manner,” said Meenesh Bhimani, M.D., co-founder and chief medical officer of Hippocratic AI. “Many studies have shown that bedside manner impacts emotional well-being and quality of outcomes.”

Photo: Pixabay