The spread of artificial intelligence in medicine is forcing hospitals to ask, who will ensure these tools improve care rather than complicate it?
Some envision a new kind of specialist, trained to analyze complex information for overburdened clinicians. Others believe the path forward lies not in inventing new titles, but in reshaping the roles that already exist.
At WakeMed in Raleigh, N.C., Neal Chawla, MD, chief medical information officer, is focused on foundations.
“All my physician and APP informaticists are getting smarter on AI with the new functionalities we are putting in, but no one is specializing in AI,” he said. Instead, he is building a tiered structure, with informaticists embedded in every specialty. “We plan to start with a solid foundation in practical and applied informatics rather than focusing on cutting-edge AI.”
For Deepti Pandita, MD, CMIO at University of California Irvine Health, training the existing workforce is the answer.
“Rather than creating entirely new roles, we believe in empowering current leaders — through targeted training and collaborative learning — to bridge the gap between algorithms and frontline care,” she said. The approach, she argued, ensures continuity and sustainability in a field that often feels defined by disruption.
Others, like Nadeem Ahmed, MD, CMIO at The Valley Health System in Paramus, NJ, are redefining the job itself. His team is restructuring the role of informaticists “from being ‘support’ driven to being more proactive in seeking solutions to ongoing clinical challenges,” he said. That means tackling alert fatigue, documentation burden and other longstanding irritants, while preparing leaders to evaluate the AI features vendors are rapidly rolling out.
“Rather than creating new informatics roles, I opt to reinvent, restructure and grow our current roles to answer the needs of today and plans for tomorrow,” he said.
But not everyone agrees the current system can handle the coming wave. Bonnie Boles, MD, RN, CMIO at Tanner Health in Carrollton, Ga., said most community hospitals lack the resources of academic centers.
“Employing teams of AI scientists isn’t feasible for us,” she said. “In our world, the best path forward is a new kind of leader already emerging: a chief AI officer who deeply understands healthcare and can safely guide how AI is integrated into clinical workflows.”
John Potts, DO, CMIO of Main Line Health in Radnor, Pa, echoed that sentiment. While he sees a “clear need for new roles like physician-algorithm specialists” — he acknowledges smaller systems may never be able to afford them. Benjamin Slovis, MD, at Temple University Hospital in Philadelphia is altogether skeptical. For him, the priority is building community and mentorship, urging junior faculty to pursue board certification and participate in governance.
“Understanding clinical workflows and acting as a bridge between clinical practice and technical solutions is the most important aspect of this role,” he said.
Some organizations are experimenting with new pathways. Adventist Health, based in Roseville, Calif., runs physician builder programs and structured mentorship while pushing for leaders with “strong data skills who can apply AI responsibly at the point of care,” said its CMIO, Amer Saati, MD, PhD.
Parkview Health in Fort Wayne, Ind., meanwhile, is letting clinicians sharpen Epic’s AI tools directly.
“In many ways, our physician builders already bridge that gap between algorithms and the front lines,” said Parkview Health CMIO Mark Mabus, MD. “The future just scales it up.”
Even in smaller settings, leaders see opportunity. At Archbold Medical Center in Thomasville, Ga., Kendall Wyatt, MD, RN, vice president of information services and chief information officer, said the focus is on two deceptively simple principles: understanding the problem and the current and future states of clinical workflow. At Cayuga Health in Ithaca, N.Y., which went live on Epic just six months ago, CMIO John-Paul Mead, MD, is already looking beyond his own role. He envisions physician and nurse informaticists, and soon provider-analysts, who can help meet the demands of value-based care.
The responses reveal no single blueprint for the future. Some systems are investing in new roles, others are doubling down on education and mentorship. What unites them is the recognition that AI is no longer abstract — and that the responsibility for making it safe and useful will fall to tomorrow’s informatics leaders.
