Dive Brief:
- Google revealed a new group of generative artificial intelligence models geared toward the healthcare sector on Wednesday, the latest in a series of developments for the new technology.
- MedLM is a family of foundation models, or systems built on broad datasets that can be used for a number of general tasks, that are “fine-tuned” for the healthcare industry, according to a blog post. They’ve been used in testing for medical documentation, pre-clinical research for drug development and helping health plan members find providers through a chatbot.
- MedLM currently includes two models, which are now available to Google Cloud customers in the U.S. through the tech giant’s Vertex AI platform. Google plans to add new models to the suite in the coming months.
Dive Insight:
Generative AI, which can be used to create new content like text or images, has become an increasingly popular topic in the healthcare sector, and tech giants like Microsoft, Oracle and Amazon have announced products geared toward the industry this year.
Though relatively few healthcare organizations have currently adopted generative AI tools, more than half of executives said they’re looking to buy or implement the products within the next year, according to a recent survey by Klas Research.
The latest family of foundation models is built on Med-PaLM 2, Google’s large language model trained on medical information.
The LLM was revealed in March this year, and the tech giant announced this summer that it would expand access to more healthcare organizations after a limited group — including HCA Healthcare, Mayo Clinic and Meditech — began testing the tools.
The two healthcare models aim to offer flexibility to healthcare organizations as they implement AI for different tasks, Yossi Matias, vice president of engineering and research at Google, and Aashima Gupta, global director of healthcare strategy and solutions at Google Cloud, wrote in the blog post.
For-profit hospital operator HCA has been using the models along with the Augmedix documentation products in emergency rooms to create medical notes drafted from conversations with patients, that can then be reviewed by physicians before they’re transferred into an electronic medical record.
The MedLM models were also tested by professional services firm Accenture to automate reading clinical documents, enrollment and claims processing; consultancy Deloitte to help health plan beneficiaries find in-network providers; and AI drug discovery company BenchSci to speed preclinical research and development.
The first of the two MedLM models is larger and designed for complex tasks, while the second model can be scaled across functions.
The tech giant added that many companies who’ve been testing MedLM are moving the tools into production or broadening their experiments.
