A $26.5 million federal bet is putting artificial intelligence in the exam room — with a goal of bringing more medical care to rural America without the price tag of staffing every clinic with specialists.
SRI, a nonprofit research institute with a Princeton, New Jersey-based team on the project, is building an AI-powered system that uses cameras, voice recognition and real-time prompts to guide nurses, nurse practitioners and general clinicians through selected medical procedures.
The pitch: A small clinic — or even a mobile medical van — could offer more treatment close to home, reducing expensive specialist staffing and long road trips to faraway hospitals.
“The biggest cost actually in delivered care is the skilled staff,” Jason Tyan, who leads SRI’s PARADIGM project, said September 14, 2026. Providing a wider range of procedures in rural areas often means hiring a multi-person specialist team or sending patients hours away, he said.
SRI received up to $26.5 million from the federal Advanced Research Projects Agency for Health, or ARPA-H, in February 2025. The money is part of the broader PARADIGM initiative, which aims to develop mobile medical units for remote communities offering services such as imaging, prenatal care, screenings and urgent treatment.
Mayo Clinic and the University of Florida are working with SRI on the effort.
A fix for the rural-care crunch?
The project arrives as rural hospitals face growing financial pressure. SRI says 100 rural hospitals have closed during the past decade, while roughly 600 others are at risk.
The financial logic is simple: let one trained general clinician do more with AI guidance, rather than requiring a specialist for every procedure.
That could mean fewer costly transfers, less time away from work for patients and fewer three-hour drives for a basic appointment.
But there’s one major catch: Nobody yet knows how much money the technology would actually save.
“It’s honestly too early for a per-visit or per-year figure,” Tyan said, adding that any hard estimate today would be “just guessing.”
SRI also declined to detail how much of the federal funding has been received or spent, saying payments are tied to project milestones and tracked by ARPA-H.
AI won’t replace doctors — at least that’s the plan
The system is designed as a digital assistant, not a substitute for medical professionals, SRI said.
Possible uses include blood draws, basic ultrasounds, splinting, lab tests, needle-based procedures and airway management, including intubation. Open surgery is not on the table.
In a typical setup, a local clinician and a small support staff could operate from a rural clinic or medical van. The AI would watch the procedure through cameras, recognize spoken instructions and provide step-by-step guidance. Remote oversight from a clinician could be available when needed.
SRI says the platform could work without a constant internet connection — a key issue in remote areas with spotty broadband.
The big risk: AI mistakes during medical care
There is no hiding the downside. AI can generate false or inaccurate information, and a bad instruction in the middle of a medical procedure could seriously harm a patient — and create a liability nightmare.
SRI says it is trying to limit that danger by pairing AI guidance with verification tools that check the system’s recommendations.
Early tests remain limited. So far, the technology has been tried on mannequins, training models and people acting as patients — not patients receiving real medical treatment.
Still, SRI said its first-year blood-draw testing produced encouraging numbers: Non-specialists using the AI guidance were accurate 97% of the time, compared with 65% without the system. More than 90% of users reported satisfaction with the blood-draw tool, according to SRI.
The project has also moved into real clinical settings at Mayo Clinic in Jacksonville, Florida, though employees played the roles of both clinicians and patients.
“Biggest saving is to the patients”
SRI is now in year two of a five-year program. By the end, the group hopes to support about a dozen essential procedures, pursue Food and Drug Administration clearance and begin formal clinical studies.
The ultimate test won’t just be whether the AI works. It will be whether insurers, Medicare, Medicaid and health systems decide AI-guided care is worth paying for.
For now, the most obvious payoff may be simpler: keeping rural patients closer to home.
“The biggest saving is to the patients because they get access close to where they are,” said Teddy Kumar, vice president of information and computing sciences and director of SRI’s Center for Vision Technologies. “They’ve got to drive three hours for a simple appointment.”
Take your tax and accounting research to the next level with Checkpoint Edge and CoCounsel. Get instant access to AI-assisted research, expert-approved answers, and cutting-edge tools like Advisory Maps and State Charts. Try it today and transform the way you work! Subscribe now and discover a smarter way to find answers.
