Technological discovery, across the centuries, tells a story of slow yet glorious evolutionary development. From the surgical hooks and probes of ancient Greece to the emerging role of AI in scientific discovery, digital innovation has charted a path of thoughtful, cumulative progress.
Today, global coalitions across health care, tech, artificial intelligence, academia and education are on distinct yet united missions to make the highest levels of expertise universally accessible in our digital age. Unlike most other industries, health care confronts a sensitive and unique dilemma: how to harness AI’s promise without compromising clinical standards and ethics.
The rural health decline: Towns like Tchula bear the cost of unprofitable hospitals and eroding health care expertise
Tchula is a place much like countless other small towns scattered across rural Mississippi—a quiet, close-knit community struggling with a worsening economic future.
As one of the nation’s lowest-income towns, jobs are few and far between, and both traffic and visitors have become rare sights in recent times. This small rural town is no exception to the predicament gripping Mississippi, where affordable health care and specialized medical services are often beyond reach. The town’s main street tells the familiar story of rural decline in this part of the country: empty storefronts, desolate suburban streets and a broken community health center that serves as the last lifeline for miles around.
The numbers also paint a sobering picture. Rural hospitals face closure threats across almost every state in the United States. A December report from the Center for Healthcare Quality and Payment Reform revealed that over half of Mississippi’s rural hospitals are at risk of closure—one of the highest rates in the country.
In the majority of states, more than a quarter of all health care facilities are at risk, while in 11 states, half or more face potential closure. The rural hospitals still serving their communities struggle not only with budget constraints but also with a technological divide that often leaves them without the sophisticated equipment, research partnerships and digital infrastructure available to major metropolitan health systems.
London and San Francisco: The twin tech titans transforming AI in health care
Urban areas are thriving thanks to AI research and innovation, but many of the most important breakthroughs never make it beyond these affluent regions—leaving rural communities behind without the training or resources they arguably need the most. When hospitals face financial pressure or closure, it’s often part of a wider pattern of regional decline, one that even the smartest AI cannot solve.
Economists and scholars identify this trend as the “cumulative disadvantage,” where the decline of core services like health care, education and employment leads to depopulation, disinvestment, falling property values and increased social disadvantage. In broader terms, the model suggests that initial disadvantages compound over time, creating a downward spiral in which communities become increasingly marginalized and struggle to recover without significant intervention.
“In cities like London, San Francisco and New York, you have the best of resources, talent and investment,” says Hatim Abdulhussein, CEO of Health Innovation Kent Surrey Sussex in the U.K. The steady concentration of capital, advanced research facilities and career opportunities in these urban centers continues to attract leading professionals, many of whom choose to remain due to the infrastructure and momentum already in place.
Abdulhussein is among a growing number of frontline clinicians working within a leading AI health hub, where emerging technologies are developed and tested in real-world settings. “We are only in the foothills. In my GP practice in northwest London, we are starting to live and breathe this,” he says. Cities like San Francisco and London offer a rare blend of academic depth and technological agility, anchored by nearby institutions like King’s College, Imperial, Stanford, and UC Berkeley, which keep them thriving.
London’s tech sector has experienced significant explosive growth, expanding its ecosystem valuation from $70 billion in 2014 to over $620 billion in 2023. This meteoric rise positions the city as Europe’s most valuable tech hub, attracting a record number of foreign direct investments, surpassing even San Francisco in tech project inflows.
Gaps in health care infrastructure and data access are stalling AI progress
A. Aldo Faisal, Ph.D., directs Imperial College’s Brain and Behavior Lab and the UKRI Center for AI in Healthcare. One of few AI engineers leading clinical trials, he combines machine learning, neuroscience and behavioral science to create human-aware AI, though he warns fragmented data landscapes and regulatory challenges are still hindering progress.
Faisal highlights a critical obstacle in the AI-friendly future: “One of the main challenges in democratizing access to AI health care technologies globally lies in the deep disparities in AI infrastructure, particularly in terms of the availability and deployability of data… and digitization of the health care system.” He says, “Only by enabling broad participation in data generation and system development can we build the high-quality, diverse datasets needed.”
Such efforts are vital because AI systems trained on comprehensive medical imaging data can detect conditions like pneumonia, diabetic retinopathy and skin lesions with accuracy that often surpasses that of specialists. “This empowers frontline health workers to make timely decisions even without on-site doctors, enabling early detection and appropriate triage,” he explains, “Most medical AI systems are trained on health care data from a small subset of a city or region with at best a few million people—this leads to quick wins, but building models that truly make AI patient-ready requires a lot more data.”
If the AI health care era arrives, its success will undoubtedly depend on data. But concerns about ethics and data use are slowing progress. Greg Slabaugh, Ph.D., professor of computer vision and AI and director of the Digital Environment Research Institute at Queen Mary University of London, observes, “A concern is that AI models trained on one demographic (urban patients) may lead to underperformance on a different population (rural patients). Efforts to produce equitable model performance are essential.”
Faisal emphasizes, “Sensitive health care decisions are those with profound implications for a patient’s physical health, mental well-being, dignity, autonomy or quality of life.” He adds, “AI could help here [and] remove unwanted variation in medical practice, making sure every patient is treated to the best scientific knowledge.”
Chatbots and automation lighten the load for frontline health care staff
“As of August 7, 2024, the U.S. Food and Drug Administration has authorized 950 AI or machine-learning-enabled devices—up from just a handful in 2015—with approvals expected to grow exponentially in the coming years,” Slabaugh explains.
AI–enabled medical devices can be extremely expensive, often ranging from tens of thousands to several million dollars, depending on their type. For example, AI-powered imaging systems can cost between $50,000 and $500,000 per unit, not including maintenance, software updates and staff training. These costs are unimaginable for many hospitals across the U.S.
AI tools, while expensive, are already proving indispensable in certain clinical settings. They screen symptoms, personalize health advice and encourage treatment adherence. Chatbots widen access and escalate emergencies to staff. By automating routine tasks, they free clinicians for more urgent work. Slabaugh points out that these technologies should be developed with the goal of supporting, not substituting, human clinicians.
Kevin Wells, Ph.D., serves as director of University of Surrey’s DataHub, specializing in applying AI technologies to solve complex health challenges. Wells notes that although the U.S. lacks universal health care coverage, its extensive and reliable mobile networks still reach deep into so-called “health care deserts.” This widespread connectivity makes telemedicine and personalized online AI support a practical and affordable solution for expanding health care access in underserved areas. While reliable mobile networks enhance telemedicine’s reach, meaningful health care improvements demand more than connectivity alone.
“In [low- to middle-income] countries and those remote/poorer communities without access to major hospital resources but with access to mobile communications services, there is major underexplored opportunity for AI to provide cost-effective support,” Wells says.
Digital assistants are changing the face of health care decision-making
For more than two decades, Jeff Pennington has helped machines make sense of the world—a veteran of over 25 years in data science and AI, with a career spanning from early work in natural language search at Ask Jeeves to leading research informatics at the Children’s Hospital of Philadelphia. He’s authored more than 20 peer-reviewed papers in biomedical informatics and data privacy. Pennington recently debuted his first book, You Teach the Machines, which explores how humans are shaping AI and its future.
Pennington has spoken extensively on how health care can integrate AI to streamline workflows, enabling frontline clinicians to care for, diagnose and treat patients using the most expert and universally available knowledge. He highlights that AI chatbots like ChatGPT have already played a transformative role in reducing health care disparities by helping patients and doctors identify conditions that might otherwise be missed. “My friend recovered from a minor shoulder injury just by using chat-based AI to diagnose and prescribe the right physical therapy. I’m not recommending that path, but it’s an example of how expert knowledge can be used right away without having to wait two weeks for an appointment,” he explains.
One area that almost always comes up in conversations about AI in health care is its potential to deliver real, practical solutions for rare and incurable diseases. These conditions affect small patient populations and historically receive little attention from the pharmaceutical industry due to limited commercial incentive. AI is helping shift this paradigm. With the ability to rapidly analyze massive datasets, identify hidden patterns in genetic information and simulate drug responses, AI tools are accelerating the pace of discovery. For all its pros and cons, AI’s potential to democratize health care is significant—especially for experts working to close longstanding gaps in care and expertise.
Even ChatGPT now features its own Medical Diagnosis Assistant—a specialized tool focused on medical knowledge that helps users understand their symptoms, receive basic diagnostic insights and get guidance on appropriate medical care. Users can describe their symptoms, upload pictures for review and explore care options. Tools like this chatbot can be very useful for determining when a health issue is urgent enough to seek hospital care. There’s also a version made specifically for practitioners, and it’s completely free to use.
Pennington has seen the challenges rural hospitals face firsthand and says the conversation around AI is shifting. “In my service on the Board of River Hospital in upstate New York, I’ve learned that cash flow is everything for rural hospitals. Their biggest risk is not having enough cash to make payroll or interest payments.”
“Many rural hospitals care for a disproportionately large number of Medicare and Medicaid-reimbursed patients,” he adds, “The radiologist workforce is aging and shrinking, and recruiting radiologists to rural health providers is close to impossible. To some extent, we have to turn to AI to augment the existing workforce just to maintain status quo.”
The integration of artificial intelligence in health care is a topic wrapped in ethical and practical dilemmas, yet its potential in serving the most deprived rural communities is certainly one to consider. Academics and tech experts all point to potential benefits, but they rely on a wider societal acceptance of AI and the expectation of data collection that makes it all happen. At this moment, the hospitals furthest from AI are those who need it most, making the challenge one of integration before necessary application.
For those struggling with rare, incurable, or complex diseases, the search for accurate diagnoses and effective treatments is both urgent and deeply personal. To them, AI is more than a technological marvel, it represents a potential lifeline. For those grappling with the challenges of rural decline and disappearing services, the urgency of this reality grows more acute. If guided by science, grounded in ethics and shaped by those who know its power best, AI could carry the promise of healing into every home—where care is no longer a privilege, but a right.
Photo by Gorodenkoff/Shutterstock
