Google is launching new healthcare AI models that aim to assist clinicians and researchers in conducting intricate studies, summarizing clinician-patient interactions and more, CNBC reported Dec. 13.
The suite of AI models, dubbed MedLM, comprises a large and a medium-sized AI model, both based on Med-PaLM 2, a language model trained on medical data initially announced by Google in March.
As of Dec. 13, the AI model is available for eligible Google Cloud customers in the U.S.
The tech company noted that the cost of the AI suite varies based on the utilization of different models, with the medium-sized model being comparatively more cost-effective.
Aashima Gupta, the global director of healthcare strategy and solutions at Google Cloud, told the publication that Google observed varying effectiveness among medically tuned AI models for specific tasks. This realization prompted the company to introduce a suite of models rather than pursuing a "one-size-fits-all" solution.
For example, Google highlighted that its larger MedLM model excels at handling complex tasks requiring extensive knowledge and significant computing power, such as conducting studies using data from an entire patient population within a healthcare organization. On the other hand, if companies require a more agile model optimized for specific or real-time functions, such as summarizing interactions between clinicians and patients, the medium-sized model is deemed more suitable, according to Mr. Gupta.
Google also said it plans to introduce healthcare-specific versions of Gemini, its latest AI model, to MedLM in the future. The tech giant released Gemini on Dec. 7, according to the publication.
Nashville, Tenn.-based HCA Healthcare is already working out plans to test the Gemini model and is working to implement MedLM.