After a long shift at UNC Hospitals, Terrell Doyle still has work to finish. Patient notes still have to be written, and charts have to be completed before he can go home. Exhaustion comes easy as the work continues and never seems to slow down.

For Doyle, a registered nurse, this reality helps explain why hospitals are searching for tools that can reduce work overload without taking away attention needed from patients. The hours spent documenting care after a long shift reflect a growing burden felt across healthcare. Healthcare systems like UNC Health and Duke Health are part of the movement to reduce the burden through the use of artificial intelligence.

Ambient AI is an artificial intelligence technology designed to listen during patient appointments and generate notes or organize patient information to reduce burnout and return more time to direct care. The goal is simple: reduce burnout caused by clinical work and give clinicians more time to focus on the care of their patients. However, hospitals are also learning that successful adoption depends on safeguards, trust and realistic expectations.

“My overall view as a nurse is that AI has potential, especially in the background parts of healthcare like documentation support, workflow improvement and data organization,” Doyle said. “But in my experience, it is not something that has yet transformed bedside care in a major way.”

In Becker’s Hospital Review, “The AI Use Case All Health Systems Are Adopting,” one survey of 43 health systems found that Ambient AI documentation tools were the only artificial intelligence application with adoption activity reported by all respondents. Fifty-three percent said they had achieved a high level of success using AI for clinical documentation.

The Becker’s Review article states that at Duke, leaders moved from experimentation to large-scale rollout after an early pilot. In January 2025, the system announced plans to expand access to the Ambient AI documentation tool to about 5,000 clinicians across more than 150 clinics after testing it with physicians and advanced practice providers.

Duke Chief Health Information Officer Eric Poon described the technology’s rapid impact in unusually strong terms in the Becker’s Review article. “I’ve been in this business 30 years, and nothing has come close to being this impactful, this quickly, at the right time,” Poon said.

Poon also said the tool changed the atmosphere of patient visits. The technology helps restore attention to interactions with the patient by reducing the need to divert focus between typing and conversation. He said Ambient AI allowed him to engage more directly with patients while the system generated a draft note in the background.

That atmosphere of less screen time and more human connection is the reason Ambient AI has gained national attention. A recent qualitative study of clinicians using Ambient AI documentation found that many reported lower cognitive burden and time savings because they didn't have to remember as many visit details at the same time as they documented them.

But clinicians say the impact can look different in daily practice. Bedside staff like Doyle are wary of exaggerated claims. He said direct patient care is built on observation, communication and judgment that AI-generated work can't replace.

“At the bedside, my job is still very human and very clinical,” Doyle said. “I’m assessing the patient, watching trends, responding to symptoms, communicating with the care team, administering medications, educating the patient and using my judgment in real time. AI has not replaced any of that.”

He said the benefits have been modest with repetitive tasks, like charting support or care-plan writing, rather than the deeper causes of burnout.

“I would describe the benefit as incremental rather than life-changing,” Doyle said. “It may make certain tasks a little faster or less tedious, but it has not, in my experience, produced a major measurable drop in burnout at the bedside.”

The lesson for hospitals may be that success depends less on buying advanced software rather than, most importantly, on how carefully it is introduced. Doyle said clinical workers must be included earlier so new systems can solve real problems instead of creating them.

The future of Ambient AI in healthcare may depend on giving time back to the people caring for patients, rather than on how impressive the software sounds.

“AI may eventually become a stronger support tool, but it still needs refinement, trust and thoughtful implementation before it can be relied on widely,” Doyle said. “The goal should not be to make healthcare less human. The goal should be to reduce unnecessary administrative burden so clinicians can spend more time caring for patients.”

Mason Spencer is a mass communications student at North Carolina Central.

Comments