It’s called Med-PaLM 2 and it’s an artificial intelligence software specialized in the medical field which, according to the Wall Street Journal, Google has been testing since last April within the structures of the Mayo Clinic research hospitals in the United States.
An embryonic stage
The overseas newspaper specifies that Google believes its updated AI model could be particularly useful in countries with more limited access to doctors. Med-PaLM 2 was trained on a series of expert demos, which could make healthcare conversations better than generalized chatbots like Bard, Bing and ChatGpt. Through a research published last May, Google showed how Med-PaLM 2 still has some accuracy problems, already glimpsed in large language models, but in the reasoning tests it gave comforting results. According to the big tech giant’s senior research director Greg Corrado, the AI model is still in its infancy. As for data processing, always fundamental in this matter, the Wall Street Journal reports that customers who test Med-PaLM 2 will be able to control their data, which will be encrypted, which Google will therefore not be able to access.