As AI floods the EHR, a quieter question is arising: What are health systems turning off?
Every major EHR vendor is in the same race right now. Epic has more than 150 AI features and enhancements in development for 2026, spanning conversational search, autonomous coding and AI-assisted charting, on top of ambient tools already in use. Oracle Health is expanding its Clinical AI Agent to automate clinical orders, including labs, imaging, prescriptions and referrals. Layered on top of all of that is a fast-growing third-party market: Ambient scribe companies have spent the past two years wiring themselves directly into Epic, Oracle Health and other EHRs.
CIOs have told Becker’s that embedding AI directly into the EHR could finally give clinicians meaningful time back.
“AI that isn’t embedded in the EHR is like having a capable assistant who sits outside the room,” Muhammad Siddiqui, CIO of Reid Health in Richmond, Ind., told Becker’s in February.
But Mr. Siddiqui also cautioned against assuming embedded automatically means better.
“If the AI is noisy, inconsistent or hard to govern, it can quietly undermine trust,” he said. “Clinicians are quick to disengage when tools feel unreliable or create rework.”
That noisy, unreliable version of embedded AI is what gets a lot less airtime: Which tools are health systems already walking back on?
According to Deepti Pandita, MD, vice president of clinical informatics, chief medical informatics and AI officer, and associate professor of medicine at Orange, Calif.-based UCI Health, chart and visit summarization is one of the tools that isn’t meeting expectations.
“Talking to my provider peers, the one feature they don’t like, despite seemingly initially being excited about it, is chart summary AI, whether released as an EHR feature or from an ambient scribe company,” Dr. Pandita told Becker’s. “The patient-level clinical summary is often not useful because, even when it reflects recentness, it is typically generated primarily from aggregated chart content (e.g., problem list/history and narrative note text) rather than from the structured encounter-level decision outputs clinicians rely on: medication changes, newly placed orders/referrals, and the specific follow-up/return plan.”
The result, according to Dr. Pandita, is a summary that sounds plausible and context-light but omits or dilutes the “what changed today” and “what we decided today” elements, so clinicians still have to reread the actual visit documentation to find the missing plan, history of present illness and assessment nuances.
This turns the summary “into extra work instead of a time-saver,” she said.
AI-driven alerts that contribute to alert fatigue are another feature that Usman Akhtar, MD, associate vice president and chief medical informatics officer at VHC Health in Arlington, Va., told Becker’s he would turn off. Alert fatigue in the EHR predates generative AI by well over a decade, and layering AI-generated recommendations on top of that existing alert burden, without redesigning when and how those alerts fire, is precisely the failure mode, according to Dr. Akhtar.
“If I could eliminate one AI-driven EHR feature, it would be any tool that turns ‘intelligence’ into an interruption,” he said. “Too often, broad AI-generated alerts or recommendations surface at the wrong time, with too little context, and add noise instead of clarity. Clinicians do not need more digital nudges; they need technology that quietly removes friction, reduces clicks, and gives them time back. AI should feel like a trusted assistant, not another pop-up demanding attention.”
Neither complaint is really about AI failing at what it was built to do. A summary that faithfully reflects the chart is doing its job; an alert that fires reliably is doing its job. The problem both chief medical informatics officers describe is a mismatch between what the tool outputs and what the clinician actually needs in that specific moment: the one line that changed since yesterday’s note, or a recommendation that shows up mid-task instead of at the point of decision.
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.