Healthcare in Georgia: An AI Lifeline

When Jim Matney felt pressure in his chest, his first thought was to reach for an antacid. But the feeling persisted, like a heavy weight that wouldn’t lift. As his wife drove him to Colquitt Regional Medical Center in Moultrie, where he is president and CEO, he imagined how embarrassed he would be if the emergency room doctor just thought it was acid reflux, told him to eat bland food and gave him a prescription.

But that’s not how the encounter unfolded in March 2017. Thanks to an early version of artificial intelligence (AI), the hospital’s EKG device detected a dangerous pattern and sent an alert: This patient is having a heart attack. An AI-enhanced lab test also quickly detected a protein released when the heart muscle is damaged.

Diagnosis and Treatment: Jim Matney, president and CEO of the Colquitt Regional Medical Center. Photo credit: Contributed

“I saw [the technician] take that EKG in front of the doctor, and I saw the doctor running back to my room,” says Matney, who was 56, with a heart-healthy lifestyle – and a family history of cholesterol plaque, or hardening of the arteries. “And then I kind of knew I was in trouble.”

Every minute is a matter of life or death when someone has a complete blockage of a major coronary artery known as STEMI, or ST-segment elevation myocardial infarction. Because AI sped up the diagnosis, within 15 minutes of stepping into the ER Matney was in an ambulance on the way to Archbold Memorial Hospital in Thomasville, 25 miles away, where he got a stent. “They got my vessel open within 60 minutes [of first arriving at the Colquitt ER]. Saved my life,” he says matter-of-factly, but still with a touch of awe. (Colquitt Regional has since opened a heart center that offers a range of cardiac care, and it began providing stent placement in June.)

That episode gave Matney, now 65, a firsthand view of the lifesaving potential of artificial intelligence. Matney is a believer in AI’s potential not only to improve patient care but also to boost efficiency and productivity, helping rural hospitals like his survive and thrive. Across the state, Georgia hospitals are tapping into AI to address a myriad of challenges, including delayed or denied insurance reimbursements, workforce shortages and patient barriers to care.

Hospitals large and small have countless data points – lab test measures, imaging results, medical notes, insurance claims, wait times, patient no-shows, to name a few. Generative AI can swiftly sift through a lengthy medical record and medical literature and guidelines to suggest a personalized treatment plan. Predictive AI can help triage patients in the emergency room, making sure people most at risk are seen quickly.

“For the person who learns how to use this, it gives them the opportunity for informed decisions that they otherwise would never have had,” says Jimmy Lewis, CEO of HomeTown Health, a Cumming, Georgia-based network of small and rural hospitals primarily in Georgia and Florida. HomeTown Health provides advocacy, consulting and education, including an Artificial Intelligence Readiness Program for Rural Health.

Aiming at Hospital Pain Points

AI is already deeply embedded in every aspect of healthcare. Every day, 40 million people pose questions about health to ChatGPT. Doctors use AI to transcribe medical notes after patient exams. The U.S. Food and Drug Administration has approved more than 1,000 AI-enabled radiology devices and hundreds more in surgery and other specialty areas. Insurance companies automate claims processing with AI. Even Medicare is using artificial intelligence in a pilot project of prior authorization, to review cases before patients receive approval for certain treatments or procedures. (The pilot program is in six states; Georgia is not one of them.)

“If hospitals aren’t using AI to check their claims before they send them in, they’re losing money.” – Jeffery Talbert, inaugural chair, Department of Artificial Intelligence and Health, Medical College of Georgia at Augusta University

So it makes sense that hospitals are racing to tap into the power of AI. But it’s a bit like paddling upstream in a rapidly moving river: They are seeking the most cost-effective AI capabilities while the technology continuously evolves. Matney went to a convention in Las Vegas to see the latest and greatest AI solutions, and he left feeling that he first needed to figure out how to use the AI capabilities he already has.

For example, he learned that lab devices don’t just produce measurements – they look for patterns in the results and suggest possible diagnoses and treatment. That helps doctors make quicker decisions. “Hospitals have probably started down this road a lot sooner than they thought,” Matney says.

Solving Problems: Ghassan AlRegib, professor of electrical engineering and AHeAD codirector and site lead at Georgia Tech. Photo credit: Kevin Garrett

While AI processes data at lightning speed, healthcare leaders often lack the information they need about how these tools function in practice to solve specific issues. Georgia Tech is one of four universities coordinating research into AI and healthcare through a new National Science Foundation-sponsored AHeAD Center, which stands for Accessible Healthcare through AI-Augmented Decisions.

The AHeAD center’s researchers conducted more than 165 interviews, including with chief information officers and clinicians at health systems, to determine their pain points – their greatest challenges and need for AI solutions. “We are aiming to evaluate, validate and guide responsible implementation of AI in health systems,” says Ghassan AlRegib, professor of electrical engineering and AHeAD codirector and site lead at Georgia Tech.

The AHeAD collaboration links health systems with technology companies and universities, providing an important platform for them to learn from each other. With funding from the NSF, universities will conduct research and provide independent analysis of the cost-effectiveness of AI-enabled devices or tools. “We truly believe that AI can solve lots of our problems in the healthcare system,” he says.

Prediction and Detection

To get a glimpse of the future that has already arrived, picture this longstanding problem and an AI solution:

An elderly patient recovering from surgery wakes up in her hospital bed, disoriented and needing to urinate. As she manages to sit up and swing a leg over the side of the bed, an alarm begins to shriek. It startles her but she doesn’t realize it was her movement that tripped it. She forges ahead, standing to go to the bathroom, but she’s off-kilter and tumbles to the floor before a nurse or aide can respond.

“We are aiming to evaluate, validate and guide responsible implementation of AI in health systems.” – Ghassan AlRegib, professor of electrical engineering and AHeAD codirector and site lead, Georgia Tech

At Emory University Hospital Midtown, a new system made by VirtuSense Technologies uses AI-enabled sensors to detect movement that predicts that a patient who is at risk of falling will try to get out of bed – 30 to 65 seconds before the actual effort. Nurses monitor the system virtually from a control center at Emory Saint Joseph’s Hospital of Atlanta and immediately ask the patient to remain in bed and wait for help – while simultaneously alerting a nurse on the patient’s floor to come to the room. (It is being extended to other Emory Healthcare hospitals.)

Falls can be life-threatening for vulnerable patients, who are often elderly and have serious medical conditions. The AI-based system led to a 20% decrease in falls and a 91% reduction in falls with injury, says Laura Fultz, chief digital information officer at Emory Healthcare.

Reducing Injury: Laura Fultz, chief digital information officer at Emory Healthcare. Photo credit: Daemon Baizan

AI predictive models also help Emory clinicians identify patients at greater risk of being readmitted to the hospital or of developing sepsis from an infection, which triggers preventive measures, she says. Emory radiologists get a boost from an AI review of thousands of imaging scans; since 2023, for example, AI found more than 2,000 potential pulmonary embolisms, or blood clots blocking an artery in a lung, Fultz says. The radiologists confirm the AI findings and interventional cardiologists move swiftly to treat or remove the clot.

Patients also may get a call from an AI voice-agent that provides instructions before surgery, keeps tabs on chronic conditions or gives basic information about medication or care. The bots extend the capabilities of Emory’s staff and allow for frequent follow-up. For example, AI called about 2,000 patients with high blood pressure who had gaps in their blood pressure monitoring and successfully closed those gaps and improved management of the condition, according to an Emory study presented at an American Heart Association conference. The patients gave high satisfaction ratings to the AI system, the study found.

“When we put together our strategy, we really put the problems first and not necessarily the technology because we want to be problem driven,” Fultz says.

An Epic-level of AI

Emory is a standout in its capacity to leverage AI. In 2025, Emory Healthcare was recognized by Epic, the most widely used electronic health record system among U.S. health systems, for being in the top 3% of customers in the use and configuration of Epic features, including AI.

While short-staffed rural hospitals could benefit from AI support, they lag in its adoption, particularly for predictive AI. “While something may be very standard in an urban hospital, like using AI to help with stroke detection and determining whether [patients] are appropriate for clot-busting drugs – that hasn’t rolled out to rural hospitals,” says Dr. Matt Lyon, an emergency medicine physician who is director of the Medical College of Georgia Center for Digital Health at Augusta University.

Helping Rural Hospitals: Dr. Matt Lyon is an emergency medicine physician and director of the Medical College of Georgia Center for Digital Health at Augusta University. Photo credit: Contributed

Unequal access isn’t the only concern. AI relies on algorithms and huge data sets, and studies show that bias creeps in if AI tools fail to account for socioeconomic differences and other disparities that affect outcomes. The use of AI also must comply with the strict federal privacy law known as HIPAA, and several states (but not Georgia) require hospitals to disclose the use of AI to patients.

ECRI, an independent, global patient safety organization, named “balancing the potential benefits and risks of artificial intelligence (AI) in clinical diagnosis” as the top patient safety concern for 2026. “Using AI diagnostic systems without strong safeguards and clinical oversight can increase the risk of missed, delayed or incorrect diagnoses,” according to ECRI’s report.

AI experts often talk about keeping “the human in the loop.” Emory Healthcare employs several systems for evaluating and monitoring the technology, and the chief AI officer is a radiologist – someone who understands the work of clinicians and the needs of patient care, Fultz notes.

AI should enhance, not restrict, the human element of healthcare, she says. “The technology doesn’t replace the humans. You still need that human interaction. … [With AI-generated phone calls,] we’re really just using it to extend our workforce to proactively reach out to patients.”

“The technology doesn’t replace the humans. You still need that human interaction. … [With AI-generated phone calls,] we’re really just using it to extend our workforce to proactively reach out to patients.” – Laura Fultz, chief digital information officer, Emory Healthcare

AI versus AI

For hospitals, the greatest AI risk may come from not using it. Increasingly, insurance companies are automating claims processing by using AI to detect coding errors or lack of documentation.

“They’re searching for things they can deny and searching for things they can put into lower buckets of care, so they have to pay less,” says Jeffery Talbert, the inaugural chair of the Department of Artificial Intelligence and Health at the Medical College of Georgia at Augusta University. “If hospitals aren’t using AI to check their claims before they send them in, they’re losing money.”

Saving Money: Jeffery Talbert, the inaugural chair of the Department of Artificial Intelligence and Health at the Medical College of Georgia at Augusta University. Photo credit: Hillary Kay

That is especially worrisome for rural hospitals, which operate on slim margins and can’t afford to miss out on payments, says Lyon, who works with rural hospitals in an MCG telehealth program. Hospitals can appeal denials and even overturn them, but it takes time. “That amount of work not only increases the expense for that rural hospital, it also delays payment,” he says.

“When you have so few resources to start with, you don’t have enough time and effort to battle some of these payment problems,” Talbert adds. “AI will be able to go through [claims] much faster to flag things and say, ‘This looks like it could be rejected or it could be downgraded’” to a lower level of care, he says.

A new state law prohibits health insurers from solely using artificial intelligence systems to make coverage decisions; it requires a human review that includes a “clinical peer.” The bill passed the General Assembly in 2026 without dissent. “We want to make sure that a person who has experience in that specialty area of care is putting their eyes on these claims so that patients are getting appropriate care,” says Anna Adams, chief government relations officer with the Georgia Hospital Association, which supported the bill.

Having a human in the loop doesn’t guarantee faster or better payment, but it offers a guardrail. Similarly, when AI is used to enhance patient care, it is a tool, not a decisionmaker.

“AI isn’t about replacing clinicians. A lot of people think that, but that’s really not what it’s about,” says Talbert. “For us, it’s really about how AI can strengthen Georgia’s hospitals by reducing burdens, expanding access and helping the rural communities survive.”

Curbing Manual Effort: Dr. Matt Zimmie serves as the vice president and chief health informatics officer for Northeast Georgia Health System in Gainesville. Photo credit: Contributed

The Doctor Will See You Now – With AI

Your doctor’s newest medical assistant might be a bot. Increasingly, artificial intelligence is embedded in the electronic health record, helping physicians focus on what’s most important to your care as AI quickly combs through your medical history and considers possible diagnoses for your current problem. It is also often used to record and summarize patient visits.

That means less burden for physicians, who can pay close attention to what you’re saying rather than double tasking by tapping into a computer.

“We’re on the leading edge of making sure that we’re utilizing this technology to reduce a lot of that manual effort,” says Dr. Matt Zimmie, a family physician who, as a specialist in clinical informatics, serves as vice president and chief health informatics officer for Northeast Georgia Health System in Gainesville.

In the past two years, NGHS has integrated 21 AI features within the Epic electronic health record system, he says. For example, DAX Copilot, a Microsoft product, acts as a scribe to draft physician notes of each patient encounter. (Doctors edit and sign off on the notes, and patients can opt for a visit without the use of AI, Zimmie says.)

Another Epic tool looks through radiology reports for mention of a lung nodule detected unexpectedly – for example, if a patient was scanned in the emergency room for a possible bone fracture after an accident. AI notifies the health system’s lung clinic to follow up on the finding.

In July, Memorial Health in Savannah was set to launch the first set of 50 AI tools through its Epic electronic health record, including support for drafting nursing care plans and hospital discharge plans. An upcoming tool called Emmie will interact directly with patients on such tasks as scheduling appointments, and it will explain the meaning of abnormal lab results, says Dr. Tim Connelly, an internal medicine physician who specializes in clinical informatics and is the associate program director at Memorial Health’s internal medicine residency program.

AI is embedded in devices, too, such as Memorial Health’s photon-counting CT scanner. Among other imaging advances it provides, it can determine if plaque in coronary arteries is fatty or calcified, which helps guide care, Connelly says.

With AI tools, “we’re helping patients, saving time and allowing for a higher quality and better quality of care,” Connelly says.