During a quick meeting^[1] —its second special session so far this year—the NIH Advisory Committee to the Director (ACD) gave unanimous approval to a “bold” $50 million program to fund a consortium to conduct research involving electronic health records (EHRs) using artificial intelligence (AI) and machine learning (ML). The goal is to seek ways to reduce health disparities. ACD members also got a preview of the agenda for the regularly scheduled meeting to be held this month.

Typically the ACD, the highest-ranking external panel advising NIH Director Francis Collins, meets twice yearly—in June and December. A May 6 meeting was called especially for the purpose of approving the consortium. Because Congress specified that funds for what NIH is jointly calling AI/ML must be used during this fiscal year (FY), which ends Sept. 30, Collins said NIH couldn’t wait until the ACD’s June 10-11 meeting.

Larry Tabak, NIH principal deputy director, gave an overview^[2] of findings by the “AI/ML Electronic Medical Records for Research Purposes” ad hoc working group to the ACD. Tabak said the current membership was a follow-up to a 2019 working group on AI.

It was not clear when the new working group was empaneled or met, but its charges were to “identify unique research opportunities for NIH to apply resources in a practical way” to EHRs, “identify EHR research challenges that AI/ML could have the greatest impact” on, and “determine barriers to the widespread use/deployment of AI/ML capabilities” that NIH support could “help overcome,” Tabak said.

The new panel made clear that, “regardless of what suite of approaches that we adopt, we’ve got to define who the partners would be that would help us scale these capabilities,” he explained, particularly “nontraditional partners,” including those who “serve the underserved” and in “marginalized parts of our society.”