[Wafer: response was truncated before the model finished its internal reasoning. Increase max_tokens, or disable thinking on this model (e.g. chat_template_kwargs.enable_thinking=false), then retry.]

Researchers at SRI International's Princeton facility are developing an AI system that guides healthcare workers through clinical procedures like blood draws, splinting and intubation when a specialist is unavailable. The technology, built with the Mayo Clinic and the University of Florida, is part of a five-year, $26.5 million federal program aimed at bringing advanced medical care to communities that lack it.

The system uses smart glasses, an overhead camera and a tablet-based interface to track each step of a procedure, confirm correct actions, flag errors and tell the user what to do next. It provides what SRI calls "reactive guidance," letting a clinician ask questions during the procedure and receive feedback in real time.

In a demonstration, Jason Tyan, principal investigator at SRI's Center for Vision Technologies, asked the system, called Amira, how to perform a blood draw. A synthesized voice recited the steps. When Tyan picked up the needle, the smart glasses projected a bright pink circle onto a patient's arm, showing precisely where to insert it.

Tyan said the spoken exchange was for demonstration purposes. In a real medical situation, a clinician would read instructions through a chatbot feature or listen through open-ear speakers in the glasses, keeping the guidance discreet.

"Right now, currently the healthcare worker, they've only been trained or specialized in certain procedures," Tyan said. "There's no way for them to cross to the different kinds of expertise. We hope we can use AI to help them perform all different kinds of procedures."

How the PARADIGM program fits

The AI guidance is one component of a larger effort called the Platform Accelerating Rural Access to Distributed and Integrated Medical Care, or PARADIGM. The program aims to create a fleet of mobile medical units that can deliver advanced services such as cancer screenings, CT scans and prenatal care to remote parts of the United States.

Millions of Americans live in areas with limited access to emergency or trauma care, sometimes called healthcare deserts. Even New Jersey, which ranks among the best states for healthcare, has regions where people struggle to get care.

SRI was awarded the funding in February 2025 by the Advanced Research Projects Agency for Health, a federal agency within the U.S. Department of Health and Human Services. The system currently covers blood draws, blood analysis, endotracheal intubation and splinting. Researchers hope to add 18 medical tasks by the end of year five.

Trust and limits of AI in medicine

AI's growing role in healthcare remains contested. Advocates say it saves time and mental energy; critics argue it threatens jobs, erodes critical thinking and creates security risks. Dr. Antonina Mitrofanova, deputy director of the Rutgers Center for Biomedical Informatics and Health Artificial Intelligence, said the truth sits in the middle.

"It is an extremely powerful tool. It can analyze things that could take us years in five minutes. And yet, we need to be very strategic about it," Mitrofanova said.

Research also shows AI helps people with prior experience the most. An MIT study published in August in Nature Medicine found that AI assistance improved the accuracy of non-experts and clinicians in diagnosing skin diseases, but non-experts tended to trust the AI-generated advice even when it was incorrect.

Mitrofanova said her center is applying for grants to teach students, post-doctorates and staff both the advantages and dangers of AI. "We need to be careful and pay extra attention to training a new generation of scientists," she said.

Why this matters for healthcare professionals

The immediate takeaway is that AI procedure guidance is in active development, and the research so far shows it helps experienced clinicians most. The MIT finding is a warning for less-experienced workers: the AI can be wrong, and the human in the room remains responsible for the outcome.

"We cannot dehumanize healthcare. We can utilize - AI is extremely powerful, it can help a lot. But we need to make sure there is a smart human in the loop," Mitrofanova said.

For healthcare workers, that suggests a practical step now: build familiarity with how AI guidance systems work before they reach your department. Training resources like the AI for Healthcare collection can help professionals understand the capabilities and limits of these tools without having to learn on the job.