Thinking About Clinical Use of AI
I was reading a recent Harvard Medical School piece on several advances in AI and medicine, and two of them particularly caught my attention.
One is CURBD, an AI framework developed by HMS researchers and collaborators that can help researchers understand how different regions of the brain communicate and influence one another. HMS describes potential applications for understanding conditions including depression, memory loss and movement disorders. Another is KinoPlex, developed in the lab of HMS cell biologist Steven Gygi, which can identify kinase activity in cancer cells and could eventually help physicians determine which kinase-inhibiting treatments may be most effective for an individual patient.
This is fascinating work, when you think about where some of these capabilities could eventually lead.
It also raises questions that I spend a lot of time thinking about in healthcare AI. How do we decide when an AI-derived finding has enough evidence behind it to begin influencing care? Who should be involved in making that decision? What changes when something developed primarily as a research capability moves into a clinical environment?
Source: Harvard Medical School, AI Framework Decodes How Brain Regions Talk to One Another, Harvard Medicine Insights, August 20, 2026.
