AI Use in Medicine is Accelerating
But Not Without Physician Concerns
Dr. José M. Zuniga, President/CEO, Fast-Track Health
Artificial intelligence (AI) is no longer a future concept in medicine; it is rapidly becoming part of everyday clinical practice. Results from an annual survey on Augmented Intelligence fielded in early 2026 show just how quickly the shift is occurring: more than four in five physician respondents (81%) reported using AI tools in their clinical practices – up from 38% in 2023.
Equally notable is the growing confidence in AI use among physicians. Seventy-eight percent of survey respondents see some advantage in using AI tools to improve the delivery of quality care. These numbers together signal a profound transformation underway in how medicine is practiced, how clinicians manage their workload, and how patients interact with the health system. But the rapid expansion of AI in healthcare also raises critical questions about governance, equity, and the future of person-centered care.
One of the most important findings from the AMA survey is how physicians are using AI today. Contrary to some public narratives, AI is not primarily replacing clinical decision-making. Instead, physicians are deploying it to assist with tasks such as clinical documentation, drafting discharge instructions, summarizing medical records, translating patient communications, and generating care plans.
These are not trivial uses given the strain on the health workforce, exacerbated during the COVID-19 pandemic. Administrative burden is one of the leading contributors to physician burnout, with many physicians spending as much time interacting with electronic health records as they do with patients. It is therefore unsurprising that 57% of physicians identified reducing administrative workload as the single greatest opportunity for AI in healthcare.
If AI can reduce paperwork and documentation demands, it has the potential to give physicians and other healthcare providers something medicine has been steadily losing for decades: time with patients. This is why the AMA prefers the term “augmented intelligence” rather than “artificial intelligence,” emphasizing that the fast-evolving technology should support, not replace, human judgment.
Despite growing enthusiasm, survey respondents said they remain cautious about AI adoption. The survey reflects a mix of optimism and concern: while the proportion of physicians whose enthusiasm outweighs their worries has increased, a significant number still report equal levels of excitement and concern about the technology, around which liability frameworks ranked high for survey respondents as an essential regulatory action to build physician trust and increase adoption of AI tools.
Among the issues most frequently raised are data privacy, medical liability, and integration with existing electronic health record systems. Most survey respondents (85%) said they also want clear feedback channels so that physicians can help shape how AI tools evolve in real-world practice. This feedback highlights a fundamental truth: successful integration of AI into healthcare will depend not only on technological innovation, but on governance, transparency, and trust.
For global health leaders, the most important question is not whether AI will transform healthcare, because it already is doing so. The real question is who benefits first and who is left behind. Historically, medical innovations tend to reach well-resourced health systems long before they reach low-resource settings or marginalized communities. If AI follows the same trajectory, it could widen existing health disparities rather than reduce them.
This is especially relevant in relation to HIV, tuberculosis, viral hepatitis, and non-communicable diseases. AI tools could dramatically improve health outcomes if deployed thoughtfully. For example, AI could help clinicians identify patients at risk of disengaging from HIV care, assist with early diagnosis of cancers or cardiovascular disease, or optimize population-level disease surveillance. But these benefits will only materialize if the technology is implemented with equity in mind.
Urban health systems, including the global Fast-Track Cities network, may play an important role in ensuring that AI strengthens—not fragments—public health responses. Cities are uniquely positioned to pilot AI-enabled solutions in real-world healthcare environments, where clinical services, community organizations, and public health departments intersect. Urban settings also offer diverse patient populations and data ecosystems that can help evaluate whether AI tools work equitably across demographic groups.
The rapid growth of AI use among physicians reflects both necessity and opportunity. Clinicians are under enormous pressure from administrative workload, workforce shortages, and rising patient needs. AI offers tools that can help relieve these pressures. But the technology alone will not determine whether this moment becomes a turning point for healthcare. Health systems, policymakers, technology developers, and civil society must work together to ensure that AI is deployed responsibly, transparently, and equitably.
The transformation of healthcare by AI is already underway. The challenge now is to ensure that this transformation strengthens – not undermines – the goals of public health: improving health outcomes, reducing inequities, and placing people at the center of care. The technology is moving fast. Our responsibility is to make sure that progress moves in the right direction.
Access the 2026 AMA Physician Survey on Augmented Intelligence report: https://www.ama-assn.org/system/files/physician-ai-sentiment-report.pdf