🎙️ Voice is AI-generated. Inconsistencies may occur.
news article
Based on facts, either observed and verified firsthand by the reporter, or reported and verified from knowledgeable sources.
In the health care industry, AI is evolving at lightning speed—and health systems aren't confident that they can keep up.
Photo description | Some health systems believe they should wait for generative AI to mature before making major investments. Meanwhile, AI developers warn that late adopters could be left behind as the tech rapidly advances.
Credit | Photo-illustration by Newsweek/Getty
In the health care industry, generative AI is growing savvier by the day. But despite the tech's rapid advancement—or because of it—some health care organizations are still deploying with caution.
Hundreds of health systems and clinics across the United States have started infusing AI into their workflows, though many of them are taking it slow, executive leaders told Newsweek in recent conversations. For example, they might pilot a generative AI tool in one department or focus group, then gather feedback and make tweaks before rolling it out to an entire specialty or provider group.
Tensions are building as tech companies urge health systems to keep pace with AI's rapid evolution—while health systems continue to pull back on the reins. As health care organizations evaluate their options, many are pondering a few key questions: Is it better to act now or wait for the next best thing? Is "good enough" AI truly good enough for now? Which is riskier: going all in or holding back?
Newsweek connected with leaders on both sides of the AI-adoption aisle. Find their perspectives below.
Even some of the most innovative hospitals are exercising caution with AI. In September, MD Anderson Cancer Center in Houston placed seventh on Newsweek's ranking of the World's Best Smart Hospitals. Yet on March 3, Dr. Lavonia Thomas, its nursing informatics officer, told Newsweek the hospital has not adopted a "great deal" of AI yet.
MD Anderson has lots of ideas on how to use AI and believes the technology has great potential, Thomas said at the health tech conference HIMSS. However, she believes that clinicians—particularly nurses—should be at the forefront of development.
Nurses bring a healthy skepticism to AI because they don't want to see their patients harmed, according to Thomas. They've also seen how the electronic health record increased administrative burden and want to ensure that new technologies will truly improve their workflows.
It's a more calculated approach, but make no mistake, Thomas said: "Nurses are not afraid of using technology."
"Nurses do have to be at the very front end, generating ideas before purchasing is even considered," she said. "[When nurses guide innovation], that's where the success is—not [when technology] is handed to them to evaluate."
Banner Health is also "pretty early" in its AI journey, Chief Technology Officer Michael Reagin told Newsweek March 13. The Phoenix-based system has been setting up governance systems before jumping into the deep end with GenAI investments. So far, it has formed a multidisciplinary committee—including ethicists, nurses and physicians—to evaluate any AI application before deployment and adopted the Coalition for Health AI's framework to evaluate and score those applications.
Still, those scorecards only show a moment in time, according to Reagin. Banner will continuously monitor all AI applications for the first six months of their deployments to ensure safe and ethical use, then will move to an annual review to confirm that "things are still stable and the model drift is not outside the bounds," he said.
Reagin characterized the strategy as more cautious than bullish, though Banner isn't shying from AI. It set a goal to reduce physicians' administrative burden by 50 percent over the next five years and believes AI can help them get there. The health system also built a general AI companion, based on the Claude family of anthropic large language models, that is currently in its pilot phase with the intention of launching next quarter.
"We're really concerned about not overbuilding and not trying to get too far out ahead of things, because of the pace [at which] these things are changing," Reagin said. "It's a little unnerving that you have an idea of what you want to do and how you want to do it, and then a month goes by, and all of a sudden, your whole perspective on life changes. And you're like, 'Wait a second, we should do it this way.'"
"Usually that happens over months or years," he added, "and now it's happening in weeks and months."
Technology developers are taking note of health systems' concerns, several leaders told Newsweek during HIMSS in early March. But generally, they disagree that slow-and-steady strategies will win the AI race.
Dr. Jackie Gerhart, a family medicine physician, chief medical officer and vice president of clinical informatics at Epic, encouraged health systems to "start now."
"Last year, 'start now' meant try AI at all—this year, it means we're moving quickly, and there are amazing things that are happening," Gerhart told Newsweek. "I worry about the 'better-than-what concept,' which means if I keep waiting for the best possible thing to come out, I am going to miss out on the amazing things I can do right now."
Dr. Aaron Neinstein, an endocrinologist and chief medical officer at the AI-driven health care operations platform Notable, offered a similar perspective. He pointed to a quote commonly used by politicians: "Don't compare me to the Almighty. Compare me to the alternative."
In other words, compare AI to the systems we have and not the systems we wish we had. The health care industry is littered with red tape and care delays, exacerbated by provider shortages and an aging population—AI doesn't have to be perfect to improve conditions for patients and caregivers, Neinstein said. While we demand 99.99 percent accuracy from AI, an exhausted physician who just worked a 10-hour shift is settling in for a few hours of admin work. It's unlikely that his work will be 99.99 percent accurate.
Trent Sanders, vice president of U.S. health care and life sciences at Kyndryl—the world's largest IT infrastructure services provider—said that in health care, "everybody's living on the fringe." According to a February Kaufman Hall report, 37 percent of hospitals are losing money; Sanders believes that some of them could see margins lift by leveraging AI to optimize workflows.
However, he acknowledged that health systems are already overburdened with technology "solutions" and could benefit from narrowing their focus.
"You can take the best advantage of AI capabilities when you have less integration requirements," Sanders told Newsweek. "That's why I'm advising if you go all in on five platforms as key to running your business, you can really start to accelerate your AI journey, your data and analytics journey, and everything that comes with it."
Both parties—health systems and solutions developers—will have to work together to achieve the best outcomes, he added.
"The advice I give to health care leaders is you can't run from [AI]; it's here, it's happening," Sanders said. "It's our job as stewards [of new technologies] to help embrace this in the right way, to drive organizational benefits so that the patient, the clinician and the health system win."
