A clinic in the US city of Mayo began testing the Google Med-PaLM 2 artificial intelligence medical chatbot in April, but an official study by the developer company itself, dated May 2023, points to a series of flaws in the system.
The Wall Street Journal has obtained internal correspondence from doctors who have tested the chatbot and Google representatives, who argue that the new model may be especially useful in countries with “limited access to doctors.”
Google Med-PaLM 2 is designed to answer questions of a medical nature and is built on the basis of PaLM 2, the language model behind Bard Google and which was officially presented by the corporation at Google I/O in May.
Because Med-PaLM 2 learns from exam questions for medical examiners, it makes the chat more knowledgeable about health issues than typical Bard, Bing, and ChatGPT. And yet, Google’s own research shows that problems with the response accuracy inherent in large language models still persist.
However, in almost all other indicators -competent reasoning, absence of errors in the comprehension and perception of the questions, decisions based on the consensus of opinions- Med-PaLM 2 shows a level of knowledge close to that of practicing physicians.
The WSJ specifically emphasizes that the data identified as part of the Med-PaLM 2 trial belongs to the clinic and directly to the clients, and is not accessible to Google.
Author:
Ekaterina Alipova
Source: RB
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
