AI and Public Health Series Part 4: How AI Can Make Healthcare More Affordable and Accessible

This series explores the ways in which artificial intelligence (AI) and machine learning (ML) can help improve health outcomes. Previous posts covered how AI and ML are improving the patient experience, addressing major causes of death and suffering, and advancing new drug discovery. Now we will examine how AI/ML systems can help address various cost drivers in the U.S. healthcare system to make healthcare more affordable and accessible.  

AI cannot solve all the inefficiencies and problems associated with our healthcare system, as deeper structural problems require multi-faceted reforms. However, AI/ML-enabled technologies and systems are already contributing toward improving healthcare access, quality, and cost. AI/ML-enabled solutions can also address waste, fraud, and abuse in healthcare administration.

America’s Expensive Healthcare System

According to research by the Peterson Center on Healthcare and KFF, the United States spends more on healthcare than any other country. In 2023, America saw expenditures of $13,432 per person—over $3,700 more than any other high-income nation. The analysis predicts that U.S. health spending will grow by an average rate of 4.6 percent annually through 2033, increasing per capita expenditures from $16,570 in 2024 to $24,200 in 2033. These costs affect healthcare access in that more than 1 in 4 adults delay or avoid needed care due to cost.

Changing societal demographics affect these costs, with the average age of Americans rising rapidly. Just 9 percent of the population was 65 or older in 1960; today, that number stands at 18 percent and is projected to rise to 23 percent over the next three decades. This “graying of America” puts pressure on policymakers to address healthcare costs as older Americans’ share of the electorate increases.

Various reports have explained how these trends threaten to push the nation into deeper and more dangerous levels of debt as healthcare becomes a major driver of federal spending.

• Data from KFF shows federal healthcare spending of $1.9 trillion in 2024, equaling 27 percent of total federal outlays for that fiscal year.

• Federal mandatory healthcare spending is growing unconstrained, with the Congressional Budget Office (CBO) estimating it will reach $1.75 trillion this year and explode to $2.95 trillion by 2035. Over that same period, Medicare expenditures are expected to grow from their current projected level of $942 billion to $1.753 trillion, whereas Medicaid spending will rise from the currently projected $656 billion to $1.025 trillion.

• Federal healthcare spending is slated to grow by 2.3 percentage points between 2025 and 2055, contributing to ever-expanding deficits and undermining the nation’s broader fiscal stability. In its long-term budget outlook, the CBO projects that federal debt will reach 156 percent of GDP by 2055 under current policies, prompting analysts to warn that a “national debt crisis” is coming.

Three Ways AI Can Help Address Healthcare Costs

AI-driven innovation can help address rising healthcare costs in both direct and indirect ways that result in significant savings. A major 2023 study, “The Potential Impact of Artificial Intelligence on Healthcare Spending,” found that “wider adoption of AI could lead to savings of 5 to 10 percent in US healthcare spending—roughly $200 billion to $360 billion annually in 2019 dollars.” Other major reports include a 2022 study by a team of 35 global medical experts titled “Economics of Artificial Intelligence in Healthcare: Diagnosis vs. Treatment” and a 2024 white paper from the Paragon Health Institute, “Lowering Health Care Costs Through AI: The Possibilities and Barriers.” 

The Paragon report identified three primary ways in which AI may help reduce medical costs: productivity gains, quality improvements, and autonomous care.

1. Productivity gains

 AI systems can help address inefficiencies in healthcare administration, potentially delivering productivity gains and cost savings. Specifically, paperwork hassles burden both patients and health professionals. A 2022 literature review of studies on waste in the U.S. healthcare system found that administrative activities account for approximately 15 to 30 percent of all healthcare spending—far higher than in other countries. For health professionals, much of this time could have instead been devoted to patient care.

AI might also help address waste and fraud in healthcare, which are significant problems for federal health spending. A June 2024 report by Republican members of the Joint Economic Committee estimated that “between $100 to $200 billion or 6 to 12 percent of Federal healthcare spending can be attributed to administrative waste.”

The report also noted that over $50 billion in improper Medicaid payments were made in fiscal year 2023, suggesting that AI could help address mistakes and fraud in both Medicaid and Medicare. Similarly, a 2024 investigation by The Wall Street Journal found that Medicare paid insurers about $50 billion for medical treatment that never actually occurred. A bipartisan bill, the Medicare Transaction Fraud Prevention Act, was introduced in both chambers to establish a pilot program for the Centers for Medicare and Medicaid to use AI to detect fraud in medical equipment purchases.

Hospitals and medical facilities are already using AI to improve drug inventory management—specifically to address “drug diversion,” or drug theft, and the costs it imposes on the healthcare system. Drug diversion occurs for a variety of reasons from personal misuse to low-level dealing. Healthcare workers are the most frequent offenders, followed by pharmacy staff, third-party vendors and contractors, and—in rare cases—patients. In a 2025 survey on new tools being used to counter drug diversion, 37 percent of respondents said they use AI/ML to identify and prevent drug diversion incidents; however, 76 percent hoped to increase their use of these tools in the future. As the resulting report noted, “Traditional methods are often insufficient to detect sophisticated schemes, and tasks that take humans hours or even days to complete can be done in a fraction of the time with the latest technologies.”

2. Quality improvements

AI systems can also lead to direct improvements in medical treatment that can potentially help lower costs. Sen. Mike Rounds (R-S.D.) has noted that, beyond their potential to achieve administrative savings for federal health insurance programs, AI/ML-enabled systems could help reduce the number of people dependent on costly procedures by identifying and treating ailments earlier.

Sen. Rounds has cited Medicare as a primary cost-driver for the healthcare system because two-thirds of recipients have at least two long-term health issues. He also said that using AI to “cure” some of these issues could save taxpayers money and potentially save lives. (Part 2 in this series identified how AI/ML is already helping to better detect and treat major ailments like cancer and heart disease.)

Another type of AI-enabled productivity enhancement that could lead to significant quality improvements is streamlining the cost and time associated with new drug design. According to various studies, it takes 10 to 15 years and around $1 billion to develop one successful drug. However, despite these significant investments, 90 percent of drugs fail in clinical trials. (Part 3 explained how scientists and drug developers are tapping AI/ML to accelerate drug discovery.)

 While there is no silver-bullet solution, some researchers predict that AI can “turn Eroom’s Law into Moore’s Law.” Moore’s Law is the historical trend of semiconductor power roughly doubling every two years while consumer costs fall; meanwhile, Eroom’s Law (Moore spelled backward) represents the trend of steadily rising drug development costs in the pharmaceutical industry. 

Here, even a small reduction could have major benefits for public health, and AI/ML is already helping to speed up discovery and reduce costs. For example, Insilico Medicine used generative AI to create a novel treatment for a terminal lung condition called idiopathic pulmonary fibrosis (IPF) that took less than 18 months and only $2.6 million to develop. Insilico was also able to move the treatment through regulatory review in 30 months—significantly faster than most new treatment approvals. IPF affects millions globally, meaning that many lives could be saved in the near term thanks to more rapid-fire drug treatments.

The head of Merck, one of the world’s biggest pharmaceutical companies, predicts huge benefits in speeding up drug research and development, thanks in large part to AI’s ability to streamline repetitive tasks so that drug developers can focus on interpreting trial results. He estimates that Merck’s new collaborations with AI health companies could accelerate the drug discovery timeline by 50 to 60 percent compared with conventional methods.

3. Autonomous care

AI-enabled quality improvements in healthcare can also bolster autonomous care, described as “the self-service delivery of a medical service via an AI system without clinician assistance.” As AI/ML takes on more tasks once exclusively performed by medical professionals or at medical facilities, it could help expand access and lower the costs of various services.

The author of Dr. Bot: Why Doctors Can Fail Us—and How AI Could Save Lives observes that, because “AI devours medical data at lightning speed, 24/7, with no sleep and no bathroom breaks,” it can find problems that doctors miss while making care more accessible. This means a chatbot could eventually become part of your medical team, offering second opinions, expanding analysis, and clarifying communications.

As noted in The Age of Scientific Wellness, “[T]hose who fold these systems into their practices will be doing their patients (and themselves) a great service” because “they are akin to having not one expert but thousands upon thousands, all working together at top speed.” For example, a new Microsoft AI Diagnostic Orchestrator correctly diagnoses up to 85 percent of New England Journal of Medicine case proceedings—over four times higher than a group of experienced physicians.

AI has already helped expand personalized care options and new medical treatments tailored to each patient’s unique needs. A recent report from the U.S. Government Accountability Office highlighted how the Department of Health and Human Services and Department of Veterans Affairs have become leading adopters of generative AI to help improve various facets of healthcare including medical research, patient monitoring, data management, diagnostics, and personalized treatment plans. More tailored care can translate into cost savings for patients and society over time.

These new tools may also help fill gaps left by a shortage of professionals. A 2024 report from the Association of American Medical Colleges projects the United States will face a shortage of between 37,800 and 124,000 physicians within the next 12 years. Recently, Mass General Brigham introduced a new AI platform to help provide patient entry services to the 15,000 hospital patients without a primary care provider.

Conclusion

Innovators and policymakers must work harder to build public trust in AI/ML technologies before more profound change can happen within America’s healthcare system. A 2023 survey by Pew Research Center found that 60 percent of Americans would feel uncomfortable if their healthcare provider relied on AI to do things like diagnose diseases and recommend treatments.

It is also important that these systems not be oversold. Even with advanced AI/ML systems in play, healthcare costs will not fall magically—many other solutions and reforms are required. Nonetheless, algorithmic technologies offer promising new ways to help improve medical treatments, expand healthcare services, and potentially create a “considerably more efficient and cost-effective health ecosystem” in the long term.