REPORT: US HOSPITALS ARE TESTING GOOGLE'S MED-PALM 2 MEDICAL AI CHATBOT

Med-PaLM 2, Google's AI chatbot for healthcare, is currently being tested in US hospitals, including the Mayo Clinic research hospital. According to the Wall Street Journal, the chatbot created to respond to medical questions is based on Google's PaLM 2, which was unveiled at Google I/O in May of this year.

The PaLM 2 language model variant designed specifically for healthcare organisations is the chatbot. The Google Bard AI chatbot, which is currently a free trial for users, is also powered by the same language model.

According to the report, the AI chatbot has been in operation since April with the purpose of responding to medical inquiries and facilitating conversations about health-related issues.

When compared to more general chatbots like Bard, Bing, and ChatGPT, the healthcare conversation capabilities of the chatbot are improved by training it on a carefully selected set of medical expert demonstrations.

Users who are testing Med-PaLM 2 will provide data, but it will be encrypted and Google won't have access to it.

Google thinks its updated model can be especially helpful in nations with "more limited access to doctors," according to an internal email obtained by WSJ.

However, compared to other doctors' responses, doctors found Med-PaLM 2's to be more inaccurate and to contain irrelevant information.

The company is still in the early stages of developing the product, according to a senior researcher who worked on the project. Despite not wanting it to be a part of his own family's "healthcare journey," senior research director Greg Corrado said he thinks Med-PaLM 2 "takes the places in healthcare where AI can be beneficial and expands them by 10-fold."