Autonomous AI is one of healthcare’s fastest-moving technologies, but some health system CIOs say they are not yet ready to let it operate without human oversight.
Becker’s asked three health system technology leaders: What technology are you watching closely but not adopting yet?
Despite leading organizations with different priorities, each pointed to autonomous, or agentic, AI. While they see significant potential for the technology, they said broader adoption depends on proving patient safety, establishing governance and resolving questions around accountability.
For Muhammad Siddiqui, CIO of Richmond, Ind.-based Reid Health, the distinction comes down to whether AI is assisting clinicians or making decisions independently.
“We’ve already seen what ambient AI can do,” he said. “But there’s a real difference between AI that drafts something for a clinician to approve and AI that acts on its own. We’ve kept a person in the loop, and we’re not ready to remove them yet.”
Before Reid Health expands its use of autonomous AI, Mr. Siddiqui said the technology must clear three hurdles: patient safety, governance and security.
“On patient safety, it has to be right every time and show its reasoning, and I need peer-reviewed evidence in a setting that looks like ours, not a vendor demo,” he said.
Dustin Hufford, senior vice president and CIO of Camden, N.J.-based Cooper University Health Care, said his organization is also closely watching agentic AI systems capable of autonomously executing multistep workflows, such as prior authorizations and patient outreach. While he believes the technology could help analysts spend more time solving complex problems and working alongside front-line staff, he said autonomous AI presents a different level of risk than generative AI tools used for summarization or content generation.
“The use of tools like ChatGPT or Gemini for summarization and augmentation carries a different risk profile, where errors, while not ideal, are generally recoverable,” he said. “In contrast, an autonomous agent acting on a flawed inference within an operational or clinical workflow may not be. That distinction shapes our approach.”
As health systems evaluate these tools, governance has become just as important as performance. Mr. Siddiqui said organizations need to know “who’s accountable and how we catch it” when AI systems are wrong, pointing to the need for “drift monitoring, bias testing on our actual population, a full audit trail, and a kill switch we control.”
Mr. Hufford said broader adoption will require enterprise guardrails, deterministic auditability and “clear and well-defined accountability structures when autonomous systems contribute to decisions that may impact patient outcomes.”
Clara Guixa, chief information and digital officer of Springfield, Mass.-based Baystate Health, also sees promise in autonomous AI but said health systems must first build the right operational foundation.
“Autonomous AI agents may deliver a blended workforce creating an intelligent ecosystem,” she said. “What is needed first [is a] defined and mature business process, vetted AI supportive care pathways, and evidence that the system fails gracefully, not silently.”
Rather than delaying adoption, all three leaders said they are preparing for it. Mr. Hufford said Cooper University Health Care is conducting controlled pilots in nonclinical workflows while developing governance models and working with legal, compliance and clinical leaders. Ms. Guixa described Baystate Health’s approach as “watch with intent,” adding that “an ‘AI formulary’ — a vetted, approved list of tools — is exactly the right framing to enterprise adoption.”
Mr. Siddiqui said Reid Health’s approach reflects caution rather than skepticism.
“I’m not skeptical of the technology. I’m disciplined about the timing,” he said. “Technology works best when it fades into the background and shows up when lives depend on it. Until these tools can do that reliably, they stay on the radar.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.