Sundar Pichai, CEO of Google, unveiled PaLM 2, a large language model (LLM) with enhanced multilingual, reasoning, and coding capabilities, at I/O 2023. The model, according to the company, can execute a wide range of activities and has been trained on 100 different languages. The AI model will be accessible in Workspace apps, Med-PaLM for medical applications, and Sec-PaLM for security, according to Pichai at the time. According to a recent report, Google has been testing an AI chatbot that can competently respond to medical-related queries.
Med-PaLM 2 has reportedly been tested since April at the Mayo Clinic research hospital, a non-profit organization based in the US, among other places, according to a report by The Verge (via The Wall Street Journal).
According to the report, Google believes the LLM will be helpful in areas with “limited access to doctors.” Greg Corrado, senior director of research at Google, stated that Med-PaLM 2 is still in its early stages.
However, Corrado stated that Med-PaLM 2 “takes the places in healthcare where AI can be beneficial and expands them by 10-fold” despite the fact that he would not want it to be a part of his own family’s “healthcare journey.”
Why does Med-PaLM perform well than other AI chatbots?
When it comes to healthcare interactions, Med-PaLM 2 will outperform more broad chatbots like Bard, Bing, and ChatGPT, according to Google. Med-PaLM is trained using a selection of demos from medical professionals.
The clients who will test Med-PaLM 2 will have control over their data, according to the report. Additionally encrypted, Google won’t have access to the data.
According to a report, Google modified its policy this week to state that it will use all publicly available data to train Bard and other AI models. Google requires data to train its AI chatbot models.