
Share
A new Elevance Health survey finds most Americans believe AI can fix healthcare's confusing paperwork and jargon, yet adoption remains low and trust thins out sharply once the stakes rise.
A new survey puts numbers on a frustration most patients already feel in their gut: healthcare is confusing, and most people don't fully understand what they're paying for or what their coverage actually does.
The poll, conducted by the Harris Poll and commissioned by Elevance Health, surveyed 2,001 adults in mid-July. The headline figures are stark. Eighty-two percent of patients said at least some part of healthcare information reads like a "foreign language." Close to half, 45%, admitted they made a healthcare decision they later regretted because they didn't understand something at the time. Only 27% said they completely understand their own health coverage.
That confusion isn't abstract. Seventy-four percent reported hitting a concrete barrier, such as struggling to parse a medical bill or make sense of their insurance terms. These aren't edge cases. They represent the median patient experience in American healthcare today.
Against that backdrop, 79% of respondents said they believe AI can improve the healthcare experience overall. That's a strong vote of confidence for a technology still finding its footing in clinical and administrative settings. Ratnakar Lavu, chief digital information officer at Elevance Health, framed the opportunity in terms of personalization. AI's ability to quickly surface and parse member-specific data, he told Fierce, allows insurers to "truly know the person" rather than treat every patient as a generic policyholder.
The expectations are concrete, not vague. Eighty-two percent of those surveyed want healthcare answers as fast as a food delivery app. Eighty-three percent want recommendations that blend their coverage, their individual health needs, and their financial situation into a single, usable answer. These are consumer-grade expectations being imported into a historically slow-moving industry.
Here is where the story gets more interesting than a simple adoption headline suggests.
Despite broad optimism about AI's potential, actual usage remains thin. Just 20% of respondents said they currently use AI regularly as a source of healthcare information. That's a wide gap between stated belief and behavior, and it's a pattern familiar to anyone who has tracked consumer AI adoption in other sectors: enthusiasm in surveys rarely translates one-to-one into daily use.

Trust narrows further as the stakes increase. Seventy-two percent said they'd trust AI for accuracy on at least one category of information, things like an explanation of benefits, a cost estimate, or medication details. But when the question gets specific, the numbers drop fast. Only 33% said they'd trust an AI platform to translate their insurance benefits. Just 30% would trust AI to estimate care costs. And when it comes to diagnosis, the number collapses to 13%.
That's a meaningful signal for any healthcare company building AI products. Patients are comfortable with AI as an information retrieval tool, something that surfaces facts quickly. They are far less comfortable with AI making judgment calls that carry financial or clinical consequences. The survey also found demographic variation worth noting: women reported lower trust in AI than men, while younger respondents were more likely than older generations to already use AI as an information source. That generational split suggests trust in these tools may shift over time, but insurers can't assume a uniform rollout will land the same way across their member base.
Lavu was candid about the implication for healthcare organizations. Tools shouldn't get embedded into products simply because the technology exists. Transparency about use cases, and clarity about where a human remains in the loop, matters more than feature velocity. "We have to implement AI at the speed of trust," he said, "which means being thoughtful about where AI adds value, not just delivering technology for technology's sake."
Elevance's own deployments reflect that caution. The company has rolled out an AI-powered tool to support utilization management reviews and speed up approvals, but it has drawn a clear line: AI is not used to make claims denials. On the member engagement side, AI is being used to identify people who might benefit from extra outreach or who are at risk of falling through gaps in care, a more proactive and arguably lower-risk application than diagnosis or benefits translation.
Lavu described the company's process as starting with listening, not technology. "We start by listening to members and identifying where healthcare is creating the most confusion or friction," he said. "From there, we determine the right solution, which may or may not involve AI."
That sequencing, problem first, tool second, is a useful discipline for any payer watching this space. The data suggests patients will reward AI that removes friction from mundane, high-volume tasks: explaining a bill, surfacing a benefit, flagging a care gap. It suggests they will resist AI making consequential judgment calls without clear human oversight, at least for now.
The survey captures a market in transition rather than a settled verdict on AI in healthcare. Belief in AI's potential (79%) dramatically outpaces current usage (20%), and trust is conditional rather than categorical: high for information retrieval, low for diagnosis. For payers and health systems investing in these tools, the data argues for incremental deployment focused on administrative friction points, explanations of benefits, cost estimates, care gap identification, rather than ambitious bets on AI-driven clinical judgment. The companies that earn trust first in the low-stakes categories will likely have more room to expand into higher-stakes use cases later. Those that move too fast into diagnosis or benefits decisions risk a consumer backlash the numbers already hint at. Patience, transparency about where humans remain involved, and a narrow focus on removing genuine friction look like the more durable strategy here, not a broad AI rollout chasing the enthusiasm captured in the topline numbers.
Tags
Original Sources
Consumers believe AI will be key in making healthcare simpler: survey
↗ https://www.fiercehealthcare.com/payers/consumers-believe-ai-will-be-key-making-healthcare-simpler-survey
About the author
Marcus began tracking AI's market implications in 2016, noticing AI-related patent filings accelerating ahead of earnings upgrades before most of the sell-side had caught on. A former fixed-income quantitative analyst, he spent two decades building models that priced risk across emerging markets before pivoting to cover the economic impact of AI full-time. His writing translates opaque technical developments into clear risk/reward terms — and he's rarely diplomatic about the gap between AI valuations and underlying fundamentals. He believes most market participants still underestimate AI's long-run deflationary effect on knowledge work.
More from The Analyst →This Week's Edition
20 November 2026
29 articles
Related Articles
Related Articles
More Stories
© 2026 Cedar & Bloom. All rights reserved.