Google has launched a set of AI models specifically designed for the healthcare industry, called MedLM. These new large-scale language models can provide medical and healthcare research services for Google Cloud customers.
MedLM AI
MedLM is developed based on Google's Med-PaLM model and its subsequent versions. This new model enhances the features of the early medical fine-tuning LLM and provides different options based on the needs and goals of the healthcare organizations using it. There are two different MedLM options, as well as many ways to customize them, but they can all answer medical questions and write medical summaries. MedLM has achieved some significant milestones, including passing the United States Medical Licensing Examination and applying it to real-world scenarios.
"The first MedLM model is larger and suitable for complex tasks. The second one is a medium-sized model that can be fine-tuned and is best for cross-task expansion. The development of these models is based on the specific needs of healthcare and life science customers, such as answering medical providers' medical questions and writing summaries," explained Yossi Matias, Vice President of Engineering and Research at Google, and Aashima Gupta, Global Director of Google Cloud Healthcare Strategy and Solutions, in a blog post. "We have tested MedLM with many companies, and now they are integrating it into their solutions or expanding their testing scope."
The official launch of MedLM will enable Google to bring LLM and its specialized intelligence to healthcare systems and life science companies. Potential and current use cases include answering doctors' questions, summarizing patient data, and analyzing diagnostic tests, whether for current patients or medical research. Google states that strict agreements govern the deployment of MedLM to ensure responsible and secure adoption of AI in the highly regulated healthcare industry. The company envisions that this technology will significantly improve patient outcomes and operator efficiency over time.
Even before MedLM, there was enthusiasm for the services it provides. For example, Mayo Clinic tested integrating Google's Med-PaLM 2 into its research hospitals as part of its broader collaboration with Google Cloud. Meanwhile, Microsoft subsidiary Nuance began integrating GPT-4 into its Dragon Ambient Intelligence platform, Amazon released AWS HealthScribe, and the UK National Health Service is driving the adoption of generative AI tools in hospitals, providing $27 million in funding. Additionally, there are ongoing pharmaceutical research projects using generative AI for drug ideation and development.
