Microsoft and Mayo Clinic partner to build specialized healthcare AI

Microsoft is joining forces with the Mayo Clinic to create a dedicated artificial intelligence model for the healthcare sector. This collaboration aims to address the growing trend of people using general AI chatbots for health inquiries, a practice that can yield unreliable or even dangerous advice due to the broad training data of mainstream models.

The new AI model will be trained exclusively on medical data, drawing from Mayo Clinic's vast collection of patient records, research, and the expertise of its clinicians. Mayo Clinic will own the model, with the ultimate goal of powering AI tools for its own healthcare providers. The organizations also plan to explore licensing the technology to other healthcare institutions and developing an AI healthcare assistant for patient engagement through the hospital's online portal.

"We needed to have the right data and the right people in the right place to be able to do this, and we firmly believe that will result in better healthcare outcomes for people who use the model," said Mayo Clinic CEO Gianrico Farrugia. The project was announced at Microsoft's Build developer event.

While the potential applications are significant, including helping patients understand diagnoses, care steps, and preventative measures, Microsoft AI CEO Mustafa Suleyman cautioned that achieving the necessary accuracy for high-stakes health questions will take "many years." Initially, the model will be tested by Mayo Clinic professionals before any broader rollout.

This venture enters a competitive landscape where other major tech companies, including Google, Anthropic, and OpenAI, are also developing AI health tools. However, Mayo Clinic's extensive experience and unique, high-quality data are expected to provide a competitive edge. The clinic has already utilized anonymized patient data to build smaller AI models for tasks like detecting heart disease and diagnosing pancreatic cancer.

“

We needed to have the right data and the right people in the right place to be able to do this, and we firmly believe that will result in better healthcare outcomes for people who use the model.