Google's Med-PaLM 2 AI Model Performs Comparable to Doctors: Almost as Good as ChatGPT?

Google's Med-PaLM 2 AI model has been tested in several hospitals since April. The tool aims to answer medical information questions and is a variant of the PaLM 2 model, which is the basic model for Google's AI platform, Bard.

According to an internal email, Google believes this tool would be particularly helpful in countries with insufficient healthcare infrastructure and limited access to doctors. Med-PaLM 2, trained by medical experts, can provide more effective medical conversations compared to general chatbots like Bard and ChatGPT.

While the model still has some accuracy issues, it performs almost as well as doctors in various aspects. Google sees potential for success in healthcare with Med-PaLM 2, despite not intending it to be part of personal healthcare. The customer data used for testing Med-PaLM 2 is encrypted with no access granted to Google. In addition to Google, OpenAI and Microsoft are also actively exploring healthcare applications.

Microsoft has partnered with medical software company Epic to develop a tool for automatically editing patient information, and their cloud service is testing OpenAI's GPT model for generating summaries and reminders from doctors' notes. 

news flash