Google Cloud, the big tech’s company’s cloud arm, is adding more organizations to test its large language model for healthcare, the company said Tuesday.
The model, named Med-PaLM 2, will be made available as a preview to an unspecified number of additional Google Cloud healthcare and life sciences customers.
Related: Amazon launching generative AI tool
Google is competing with Microsoft, Amazon and other companies looking to grab market share as more healthcare organizations seek to adopt generative artificial intelligence. Released in March, Google’s healthcare-focused generative AI model scored 86.5% on doctor licensing exams and is tailored for medical questions.
Users of Google's generative AI model have included Nashville-based HCA Healthcare, Rochester, Minnesota-based Mayo Clinic, Bayer Pharmaceuticals and electronic health record vendor Meditech.
“Right now, it’s early days in determining which [large language model] model works for which use case,” said Aashima Gupta, global director for healthcare strategy and solutions at Google Cloud. “It’s essential to have deep partnerships and deep collaboration. That’s our approach. That's how Google wins.”
Google's strategy is to be deliberate with its rollout of generative AI for healthcare organizations, said Greg Corrado, the company’s head of health AI. The company is looking to do it safely and effectively, he said.
“It’s a little bit of a naive view to imagine that large language model systems are able to do anything expertly off the shelf,” Corrado said. “These systems always require identification of specific use cases and then further development to optimize quality and performance for those use cases.”
Google isn’t the only big tech firm looking to disrupt healthcare with generative AI. Amazon launched a generative AI tool in July focused on clinical documentation. Microsoft is working with subsidiary Nuance Communications and EHR giant Epic Systems to make OpenAI's language model GPT-4 capabilities available for health systems.