The Wall Street Journal reportedIt is said that Google recently announced the artificial intelligence model "Med-PaLM 2" designed for medical applications, which has been tested in institutions such as the Mayo Clinic since April this year.
"Med-PaLM 2" is a derivative design of the "Med-PaLM" proposed earlier, which is based on the design officially proposed by Google this year.Large Natural Language Model "PaLM 2"It is also the large-scale natural language model behind the "Bard" service, which can understand the user's spoken content and quickly take corresponding actions.
According to internal emails, Google expects "Med-PaLM 2" to have better information understanding and application effects in the medical system than its own "Bard", competitor Microsoft's Bing, and OpenAI's ChatGPT.
However, because current large-scale natural language models still encounter information inaccuracies when interacting with artificial intelligence technologies, the application of technologies like "Med-PaLM 2" in the medical system may cause considerable skepticism and panic. After all, misjudgments can lead to subsequent medical misdiagnosis and even worsen a patient's condition.
However, if we consider that doctors themselves can make misdiagnoses during the diagnostic process, the errors in medical judgments driven by artificial intelligence "Med-PaLM 2" seem to be relatively reduced, but it may still take more time to verify.
In addition, the report stated that when the medical system tests and uses "Med-PaLM 2", all data content will be encrypted, and even Google will not be able to access such data.
Currently, "Med-PaLM 2" is still in its early stages of development. According to Google's previous statement, it has no plans to replace the jobs of existing medical personnel, but it is expected to significantly improve medical efficiency.
