by
John R. Fischer, Senior Reporter | June 13, 2023
Mayo Clinic will pilot a new Google Cloud Generative AI technology. (Photo courtesy of Mayo Clinic)
Mayo Clinic is test-driving a new generative AI solution designed by Google Cloud to create relevant content that will personalize care experiences using data from different formats and disparate sources.
On June 6, Google Cloud introduced Search Enterprise on Generative AI App Builder, an application that creates chatbots equipped with Google's search engine capabilities and generative foundation models, to pull structured and unstructured data from various documents, databases and intranets, based on specific information searches initiated by users.
The search giant, which
entered a 10-year partnership with Mayo Clinic in 2019, will use feedback from these tests at Mayo Clinic and other early adopters to refine the capabilities of its Gen App Builder for use in clinical settings before distributing it to other healthcare providers,
according to CNBC.
An expedient and flexible tool
Generative AI uses neural networks to identify and utilize patterns in existing data from diverse sources to create new and original text, images, animations, 3D models, and other types of content. It is often used to solve complex problems, and has increasingly become the subject of scientific research into the potential role it could play in healthcare.
"We expect enterprise search to help with a wide variety of things, such as making it easier to search on patient medical histories, more user-friendly provider directories that take advantage of natural language search instead of a rigid keyword search, and things like clinical trials search to speed up the process of identifying participants, and to find information on clinical trials underway," Vish Anantraman, chief technology officer of Mayo Clinic, told HCB News.
Gen App Builder is designed to learn and become more selective in its search inquiries. For instance, it allows customers to review responses and eliminate biased and unverified content, actions that the solution incorporates into future information searches.
Recognizing limitations
The accuracy of generative AI comes down to the specific outputs of data that models are based in; meaning that the chance for errors rises when such technologies process new information that they have no prior training around to verify if facts are correct.
