AI in healthcare: Google's Med-PaLM 2 chatbot enters testing phase in US hospitals

economictimes.indiatimes.com:

AI healthcare realm has taken a significant leap with its Med-PaLM 2 tool now having entered the testing phase at the Mayo Clinic research hospital in the US. This technology, an enhanced version of the PaLM 2 AI language model introduced at this year's Google I/O, aims to answer medical queries and facilitate conversations on health-related matters. According toThe Wall Street Journal, the tool has been in use since April and is expected to be especially valuable in regions with limited access to healthcare professionals. Med-PaLM 2 is the result of training the language model on a curated dataset of medical expert demonstrations, which Google believes will enable it to outperform generalised chatbots like Bard, Bing, and ChatGPT when engaging in healthcare conversations. Google's research, which the company made public in May, was also mentioned by The Verge, showing that Med-PaLM 2 still suffers from some of the accuracy issues which generally persist in large language models.

Google training Bard with scraped web data? Here’s everything you may want to know

AI ChatGPT-powered smart toys may soon be unveiled; Here’s everything parents need to know ahead of holiday season

Google's AI in Gmail gets it right with ‘Help me write’ feature

While the accuracy of these large language models remains a major concern, Google's Med-PaLM 2 performed comparably to human doctors in areas such as reasoning, consensus-supported answers, and correct comprehension. The use of AI in medical settings has raised privacy concerns, prompting Google to ensure that customers testing Med-PaLM 2 retain control over their encrypted data, shielding it from the tech giant’s access.Mashable has shared that Google recently