How Three Harvard Medical School Experts See AI Reshaping Health Care Delivery 

AI is reshaping medicine by enhancing care, safety, and education. Experts stress thoughtful use, equity, and human judgment to ensure technology augments, not replaces, clinicians.

A hand tapping a screen with a crimson overlay.

Artificial intelligence (AI) is no longer a futuristic concept in medicine. It is already reshaping how health care professionals deliver care, manage operations, and prepare the next generation of clinicians. From radiology reading rooms to nursing units and clinical encounters, AI is improving outcomes, streamlining workflows, and enhancing patient experiences. At the same time, experts at Harvard Medical School (HMS) emphasize that thoughtful integration and careful safeguards are essential to ensure that AI delivers on its promise without compromising safety, equity, or human judgment.

Transforming the Patient–Clinician Experience

One of the most visible and immediate applications of AI has been in the clinical encounter itself. Roger Daglius Dias, MD, PhD, MBA, faculty director of HMS’s Leading AI Innovation in Health Care blended certificate program, points to the growing use of “AI scribes” that rely on ambient listening to document visits. “This has led to improved patient experience, since doctors can better communicate with patients, and has also led to a reduction in clinicians’ burnout and the time spent on documentation after clinic hours,” he explains.

For Dias, this is just the beginning. The first wave of medical AI has focused largely on reducing administrative burden, but the next wave will move to the bedside. “AI has huge potential to reduce medical errors and improve patient safety and the quality of care,” he says.

Redefining Clinical Roles in Imaging

Radiology is another field where AI is already making a measurable difference. Pranav Rajpurkar, PhD, assistant professor of biomedical informatics at HMS and faculty lead for two HMX AI in Medicine online courses covering Medical Image Interpretation and Biomedical Signal Interpretation, notes that AI is helping radiologists identify critical findings, accelerate reporting, and reduce backlogs. These tools, he explains, “allow radiologists to deliver faster and more reliable results, particularly in acute care settings where every minute counts.”

Rajpurkar believes that the real shift is in how AI will redefine professional roles. “Rather than the old framing of ‘AI assists clinicians,’ I see the field moving toward a clearer division of labor. AI is poised to autonomously handle high-volume, low-complexity tasks such as normal radiographs or screening exams while clinicians focus on more complex and ambiguous cases.” Foundation models trained on millions of medical images, he adds, are especially promising because they can adapt to a broad range of findings, reflecting the realities of clinical practice far more closely than today’s narrow, single-use algorithms.

Enhancing Patient Safety and Nursing Practice

For Debra Blyth-Wilk, DNP, JD, RN, CPHQ, capstone director for the Safety, Quality, Informatics, and Leadership certificate program, the promise of AI lies in making care “more reliable, safer, and more equitable.” In her work at the intersection of patient safety, quality, and nursing, she has seen AI’s impact across many dimensions of care. Predictive analytics help clinicians anticipate risks, such as sepsis, falls, and readmissions, before they occur. Natural language processing is also uncovering hidden risks in safety reports.

“Nurses are on the front lines of patient safety, and AI can make their work more reliable and less dependent on memory or vigilance alone,” Blyth-Wilk explains.

She also sees AI transforming education. Adaptive platforms can personalize learning at scale, while generative AI can create realistic patient scenarios for simulation training. “Our challenge as educators is to prepare students for a future where safety and quality will increasingly be supported by digital tools,” says Blyth-Wilk. Still, she cautions, “AI is a tool to augment human judgment, not replace it. Digital literacy and ethical awareness will be critical.”

Cautions and Considerations

While the opportunities are enormous, the faculty agree that implementation must be deliberate. Dias stresses that clinicians must never lose sight of their judgment: “Providers need to understand when to trust AI recommendations and when to override them.” He also warns that poor data quality or unrepresentative training sets can risk embedding bias and worsening disparities.

Rajpurkar emphasizes the need for clarity around responsibility. As AI takes on more work, safe adoption will depend on workflows that specify when algorithms can act independently and when human oversight is essential.

Blyth-Wilk underscores the cultural risks of overreliance: “Algorithms can embed bias, predictions can mislead, and automation can lull us into complacency.” For her, the solution lies in preparing clinicians and students to engage critically with technology, balancing efficiency with ethical awareness.

Looking Ahead

Despite the cautions, the faculty envision a future of opportunity. Dias highlights multimodal AI systems that can integrate imaging, lab values, genomic data, and clinical notes, offering more complete insights and supporting truly personalized medicine. Rajpurkar points to generalist AI models with broad clinical reasoning capabilities, which could help expand access to diagnostic expertise in regions with limited specialists. Blyth-Wilk sees virtual nursing models that combine AI-enabled monitoring with human care, expanding reach while preserving the personal connection at the heart of nursing.

Together, these perspectives show a clear trajectory: AI will become a partner in care through amplifying human expertise, reducing risk, and unlocking new models of delivery. However, success will depend on the choices that health care leaders make today, including how they design workflows, address bias, educate clinicians, and keep patients at the center of innovation.

As Blyth-Wilk reflects, “The promise of AI is not in the algorithms themselves, but in how we harness them to make care safer, more equitable, and more reliable.” Her words capture both the caution and the potential. AI will not replace the art of medicine, but with careful stewardship, it may help clinicians write its next chapter.