Google launches MedLM, an AI doctor that will be used in healthcare

Published Dec 17, 2023

No headings found

In a recent announcement, Google introduced MedLM, a set of artificial intelligence (AI) models tailored specifically for the healthcare industry. This model is built on Google's second iteration of medically tuned AI tools which previously sparked news on the internet. In this article, let us dive into discussing more about the latest Google MedLM in more detail.

Google MedLM: Overview

Med-PaLM 2: Accuracy?

The Google MedLM is based on Google's earlier research projects of Med-PaLM 2 , which had achieved an impressive 85% accuracy on U.S. medical licensing exam questions, reaching what Google equates to an "expert" doctor level. This marks an 18% improvement over its predecessor, positioning Med-PaLM 2 as a significant advancement in medically tuned AI.

Google MedLM Models

Google's MedLM consists of two models, each designed for specific tasks based on their sizes. The first model, larger and suited for complex tasks, contrasts with the second, a medium-sized model ideal for fine-tuning and scalability across various healthcare-related applications.

Future Expansion with Gemini

Looking ahead, Google plans to integrate Gemini-based models into the MedLM family, enhancing its AI capabilities. Furthermore, this strategic move aims to further expand MedLM's applications within the healthcare sector in the coming months.

Real-world Applications

MedLM is positioned for use across the healthcare industry, including applications in hospitals, drug development, and patient-facing chatbots. In addition, several organizations have already tested the suite, with promising results.

HCA Healthcare utilized MedLM to augment technology from Augmedix, improving the quality and scalability of transcription technology in hospitals.
BenchSci incorporated MedLM into its R&D platform, enhancing the ASCEND engine's ability to sift through clinical data for drug discovery.
Accenture and Deloitte joined as early adopters, employing MedLM to speed up AI-powered processing of clinical documents and improve patient-provider matching for contact center agents.

Collaboration and Adoption

Collaborating with clinical documentation startup Augmedix and drug research company BenchSci, Google has successfully implemented MedLM to convert data into medical notes and expedite drug development processes. In addition, partnerships with Accenture and Deloitte aim to accelerate enterprise adoption and improve healthcare organizations' information retrieval.

Also Read

Google MedLM: Availability and Conclusion

Google's MedLM marks a significant leap forward in AI tailored for the healthcare industry. With its specialized models, real-world applications, and strategic partnerships, MedLM holds the potential to revolutionize how AI contributes to various healthcare domains. However, the Google MedLM is only available for the U.S. Google Cloud customers via the Vertex AI platform, and its availability outside of the United States remains uncertain.