Google’s medical chatbot is being tested by hospitals

Google’s Med-PaLM 2 artificial intelligence model, designed to answer medical questions, has been in testing at the Mayo Clinic research center since April. Med-PaLM 2 is based on the PaLM 2 model that Google introduced at this year’s I/O 2023 conference and is based on Bard.

Google believes its updated model could be particularly useful in countries with “more limited access to doctors,” the Wall Street Journal said in a report based on an internal Google email. Med-PaLM 2 is trained on a set of specialized medical data, which Google says will perform better than generic chatbots such as ChatGPT and Bard at answering medical questions.

Google medical chatbot

In this report, it is also said that Google’s Med-PaLM 2 model is still not accurate enough and has problems. However, Med-PaLM 2 performs more or less as well as real doctors.

The Wall Street Journal also noted that Med-PaLM 2 encrypted data is controlled by the customers themselves and Google will not have access to it.

Greg Corrado, Google’s chief research officer, explained to this publication that Med-PaLM 2 is still in its early stages of development. Corrado also mentioned during his talk that while he prefers not to use the tool for his family, he believes that Med-PaLM 2 can “cover situations in healthcare where AI can help.” be useful.”

According to Google’s explanation at the I/O conference, the PaLM 2 model supports more than 100 languages, can code and even find code errors, has stronger reasoning power, is faster, and is supposed to be used on 25 internal Google products. . This model is an updated version of the original PaLM model that Google introduced in 2022.