Google is testing an AI chatbot in the US that can answer medical questions. Over time, the chatbot will also benefit countries that have limited access to doctors.

Google is hoping for a breakthrough in the medical field with its chatbot. The Med-PaLM 2 AI bot is already being tested in several US hospitals. The Wall Street Journal knows that. Med-PaLM 2 is a variant of PaLM 2 that forms the basis of ChatGPT competitor Bard.

Better trained

The model is trained using extra curated medical data. According to Google, this makes it easier to hold discussions about medical topics. However, preliminary research shows that even the medical chatbot regularly misses its target, is less accurate, and provides less relevant answers than real doctors. In general, however, the performance of the bot is on a decent level.

The chatbot would not collect any patient data, and neither would Google have access to it. While the bot is still in the early stages of development, there seems to be a goal to adopt the technology more widely in the long term. It won’t be instantaneous for us. Bard has not yet launched in Europe because Google has not yet been able to adapt the bot to European rules on data, AI and privacy. An excursion of such a bone into the medical world certainly seems impossible here.