Latest In

News

Testing Of Google's Medical Artificial Intelligence Chatbot Underway In Hospitals

Testing of Google's medical artificial intelligence chatbot underway in Hospitals since April. The artificial intelligence tool is known as Med-PaLM 2 and it been tested at research facilities such as the Mayo Clinic, as stated in a recent article published by The Wall Street Journal.

Daisy-Mae Schmitt
Jul 10, 2023253 Shares84460 Views
Testing of Google's medical artificial intelligence chatbot underway in hospitalssince April. The artificial intelligence tool is known as Med-PaLM 2 and it been tested at research facilities such as the Mayo Clinic, as stated in a recent article published by The Wall Street Journal.
PaLM 2, which was shown at the Google I/O conference in May, has been upgraded to become Med-PaLM 2, which was also presented at that conference. Google's Bard is based on a language model known as PaLM 2, which was developed by Google.
According to the article, Google's conviction that the modified model could be particularly beneficial in countries where access to doctors is limited was revealed in an internal email, and it is mentioned in the article. Google believes that by training Med-PaLM 2 with a carefully selected collection of medical expert demonstrations, it will be able to surpass conventional chatbots like as Bard, Bing, and ChatGPT in conversations pertaining to healthcare.

Groundbreaking research in health AI | The Check Up ‘23 | Google Health

According to an internal email that was obtained and published by The Wall Street Journal, Med-PaLM 2 has the ability to not only answer medical questions, but also summarize enormous amounts of health data and documents, as well as provide reminders.
The Wall Street Journal reports that despite the fact that Google's Med-PaLM 2 is a promising AI tool for medical information, it still demonstrates accuracy concerns that are comparable to those that are observed in other big language models. When compared to responses from other doctors, the answers produced by Med-PaLM and Med-PaLM 2 were found to contain a greater number of errors and irrelevant details by medical professionals in a research that was made public by Google in the month of May.
However, Med-PaLM 2 displayed performance that was equivalent to that of real doctors in a number of other criteria, including demonstrating evidence of reasoning, offering responses that were supported by consensus, and demonstrating proper comprehension.
In addition, the article explains that clients that take part in the Med-PaLM 2 testing will have control over their encrypted data, preventing Google from gaining access to it.

Conclusion

According to a article from The Wall Street Journal, Greg Corrado, senior research director at Google, admitted that the development of Med-PaLM 2 is still in its preliminary phases. Corrado shared his concerns over the possibility of implementing Med-PaLM 2 into the medical care provided to his own family. However, he is of the opinion that the AI tool had the potential to dramatically amplify the areas in healthcare in which AI can be useful, hence expanding their influence by a factor of ten.
Jump to
Latest Articles
Popular Articles