Physicians are already deep into artificial intelligence adoption—67% of doctors use AI tools daily in their practice and nearly 90% use AI at least weekly.
Clinicians are using AI daily for writing notes, drafting documents, patient communication and simplifying routine tasks, according to a recent survey from Offcall, a physician-founded company that provides salary and work transparency, along with financial education for doctors. Offcall surveyed 1,000 physicians across 106 specialties in October to understand how they are using AI tools.
Eighty-four percent of doctors say AI makes them better at their jobs, and 78% are optimistic about the future, believing it will make patients healthier, the confidential poll found. Only 3% of doctors say they never use AI.
The survey reveals widespread, organic adoption that contrasts sharply with the sluggish pace of formal institutional rollouts, according to Offcall.
“Based on these numbers, I think [AI is] being used more than physicians are talking about publicly, and it’s only going to continue to grow,” Graham Walker, M.D., Offcall co-founder, said during a podcast episode discussing the survey results.
Most physicians—81%—say they are frustrated with their organizations’ approach to AI tools and the speed of AI adoption. That dissatisfaction comes down to three things—speed, influence and communication. There’s a growing misalignment between physician autonomy and organization decision-making, according to the survey results.
Doctors feel that they don’t have a seat at the table when it comes to deploying AI in their organizations. Nearly three-fourths of physicians (71%) say they have little to no influence on which AI tools get used. Thirty-five percent said their opinion is considered but they have no direct influence, while 67% said more influence in these decisions would increase their job satisfaction.
Only 10%, mostly private practice physicians, say they have any sense of real control over AI decisions.
Related
“Physicians want a seat at the AI table. They want to be heard. I don’t think any physician is going to say that the one cardiologist or the pediatrician should be the one that decides which tool everybody else uses,” Walker said. “I think the physician also would want a tool that has been vetted or tested or works with their population.”
This disconnect between clinicians and administrators mirrors the early days of the electronic health record (EHR) system rollout, according to Walker.
The HITECH Act in 2009 significantly boosted hospital adoption of EHRs by offering substantial federal incentives.
“The challenge, I think, looking back, is that the EHR tools were not necessarily designed with the end user in mind. They were not designed to make a nurse have a delightful time using the electronic health record similarly with a social worker, a pharmacist, a doctor, that was not the primary goal, because that’s not what the law said. The law had these meaningful use criteria that required that the EHR did these regulatory check boxes,” Walker said during the podcast interview.
“The parallel I see there is AI will get adopted to us, not with us, unless we are actively speaking about it and speaking our minds and saying how we need this stuff to work,” he said.
Nearly half of physicians rate their organizations’ communication about AI as poor.
When asked about their organizations’ speed of AI adoption, one physician said, “Not good, we are not adopting AI nearly fast enough. Meanwhile, I’m using ChatGPT and OpenEvidence on my own.”
Another physician said, “Hospital keeps saying ‘We’re evaluating AI tools’ for the past two years. Nothing happens.”
Physicians’ biggest concern about AI isn’t that these tools will replace them but misaligned incentives, the survey found. Doctors fear that AI, placed in the hands of payers or administrators rather than physicians and healthcare teams, could undermine patient care.
“All technologies are arguably neutral,” Walker said, but the incentives behind it are not. A payer-controlled model that nudges toward fewer admissions or fewer tests poses a fundamentally different risk than a physician-centered tool designed to improve care.
Related
“I think that physicians are in the unique position of, hopefully, being less biased. We’ve taken the Hippocratic Oath. When you compare that to a payer or a health system that has a very hard financial benefit that’s arguably attached to getting care or not getting care, getting them into the hospital or not getting them into the hospital. I think the patient and the physician feel way closer to what the ‘right answer’ is,” Walker noted.
Misaligned incentives and payer or admin control of AI tools was cited as a top fear by physicians (30%), followed by liability and litigation (20%), loss of medical art/analytical thinking (15%), technology failures/reliability (15%) and reduced human connection (10%).
Among the surveyed physicians, 80% cited HIPAA security as a major concern when adopting new AI platforms.
Asked about top AI use cases and priorities, physicians overwhelmingly want documentation and administrative relief. Nearly two-thirds (65%) cited documentation and scribing, followed by administrative burdens (48%), then clinical decision support (43%), AI to augment capacity (27%) and data aggregation (20%).
The top 5 AI tools used by physicians are OpenEvidence (45%), ChatGPT (16%), Abridge (5%), Claude (3%) and DAX Copilot (2.4%). Doximity was cited by 1.8% of physicians as was Epic AI.
Going forward, healthcare organization leaders need to earn physicians’ trust and buy-in and they can do that by moving faster and adopting the tools physicians are already using, giving doctors more influence in decision-making and communicating transparently, Walker said.
