At , we have built a culture that meets challenges boldly and treats obstacles as opportunities. Artificial intelligence presents just such a challenge. The question is no longer whether AI will become part of academic medicine. It already has.
The real question for the , and 返字心頭 Health more broadly, is whether we will shape its use thoughtfully and responsibly or simply be shaped by it. Thus, our responsibility as an academic health system is to determine how AI can strengthen our three missions of patient care, education, and research, while preserving the human values that define us.
Across the country, many research-intensive medical schools and their teaching hospitals, including ours, are approaching this question systematically. They are , investing in secure clinical data environments, integrating AI literacy into medical education and creating centralized computational resources that allow clinicians and scientists to .
Patient Care
The earliest impact of AI in clinical medicine will not be the replacement of providers but rather a reduction of friction in everyday clinical work.
Many of you already use tools such as to draft clinical notes during patient encounters, while Epic tools such as In-Basket ART message triage and Epic Insights chart summarization help our clinicians navigate increasingly busy and complex care. Early experience suggests that these technologies can return meaningful time (i.e., two hours a day) to clinicians by reducing documentation burden.
At 返字心頭 Health, we are already seeing how carefully targeted AI deployments can improve patient access. In just a few months, our call-center AI voice agent Aimee has handled over 600,000 patient calls, reducing abandonment by 56% and increasing scheduled visits by 21%, an improvement that would otherwise have required 40 additional staff. Aimee is now expanding her capabilities to include appointment rescheduling, cancellations and other patient interactions.
Artificial intelligence is also increasingly embedded in diagnostic specialties. In radiology, we are using a product called Eon that uses AI to find incidental findings in reports and help with care coordination for follow up of lung nodules and pancreatic findings. Across the country, to detect intracranial hemorrhage, breast lesions and fractures while helping prioritize urgent studies and standardizing reports. Over the next few years, these systems will generate structured quantitative analyses, measuring tumor volumes, coronary calcium, emphysema burden and liver fat, with radiologists validating the findings. AI also has the potential to exploit routine tools in unexpected ways; undergoing screening mammography, AI-quantified breast arterial calcification levels showed a strong doseresponse relationship with future cardiovascular events, independent of traditional risk factors.
返字心頭's Dr. Yael Bensoussan is co-leading a multi-site NIH study using AI to determine whether the human voice can serve as a biomarker for diagnosing disease.
But we should be clear-eyed about risks. Technology that improves efficiency can also erode vigilance if used uncritically. suggested that prolonged exposure to AI-assisted colonoscopy tools may reduce adenoma detection rates by 20% when the technology is removed, raising concerns about skill deterioration. In other words, AI must function as a clinical co-pilot, not an autopilot. We must have safeguards in place to ensure that our AI tools are used ethically and perform accurately through use of continuous monitoring systems that can detect AI hallucinations, model drift and provider de-skilling.
Over the next year, our clinical strategy will focus on disciplined integration of other AI tools directly into clinical workflows in partnership with Tampa General Hospital. The goal is not technology for its own sake, but measurable improvements in patient care quality, safety, access and efficiency. The recent rewiring of TGHs Care Coordination Center (C3) using Palantirs technology combines local clinical and operational expertise with cutting-edge AI to optimize patient flow through tools like the Care Progression Navigator and patient safety enhancements through continuous monitoring for sepsis, heart failure and clinical deterioration. Through deliberate investment in Epics Generative AI capabilities, 返字心頭 Health-TGH has earned a place among the top decile of AI adopters globally and distinguished itself as the first in the nation to bring Epics conversational search to life, enabling clinicians to interact with Epic and clinical data the way they would interact with ChatGPT.
In the next few years, the very nature of patient care will evolve. We will see more states allow AI algorithms to prescribe medications without a physician in the loop, such as . Technology companies are investing heavily in that combine virtual visits, prescription management, automated triage and personalized health guidance. Patients will increasingly expect health care to be responsive, convenient and continuously available. But without seamless interoperability with the patients medical record and communication with providers, such commercial services could exacerbate care fragmentation and waste.
Academic health systems must meet rising patient service expectations while preserving continuity of care and patient trust. If we succeed, AI will restore time for clinicians and strengthen the patient-physician relationship. If we fail, it will simply add cost and complexity to an already strained system.
Education
Artificial intelligence is already present in medical education, whether we formally acknowledge it or not. I use OpenEvidence multiple times a day, and I suspect that is true of many of our providers and learners. For the latter, AI will help explore clinical questions, generate explanations, summarize material, simulate cases and provide instant feedback. In an age when medical knowledge may be doubling monthly, AI is now a necessity for self-directed and life-long learning.
The challenge for medical schools is not whether AI will be used, but how it should be used responsibly. Professional organizations and academic leaders have begun outlining principles for . These emphasize transparency about when AI is used, protection of patient data, equitable access to educational tools and continuous evaluation of outcomes.
At 返字心頭 Health MCOM, we are beginning to address this systematically. We have integrated AI training into our evidence-based clinical reasoning courses. Over the next year, we will introduce AI competency-based objectives into our mapped, integrated undergraduate medical education (UME) curriculum. This will include pilot activities and competency frameworks designed to ensure that our graduates are not only AI-literate but capable of using AI critically in clinical reasoning and decision-making. In parallel, we will expand AI training across graduate medical education (GME) and continuing professional development to help practicing clinicians learn how to use these tools safely and effectively.
Looking ahead, AI will reshape how medical expertise itself is defined. Historically, medical education has emphasized mastery of vast bodies of knowledge. AI will increasingly make that information instantly accessible. Physicians who thrive in this environment will be those who excel in judgment, critical thinking, communication, ethics, leadership and the supervision of complex clinical systems.
Our responsibility is to prepare current and future physicians and other providers who can work with AI without allowing it to erode the intellectual discipline and compassion that define the health profession.
Research
Among our three missions, research may ultimately experience the most profound transformation from AI. Even today, AI tools are accelerating many elements of scientific work. Acting as transparent, human-supervised tools, AI can aid in reviewing literature, simulating outcomes, identifying research cohorts, generating analytic code and seeking connections among a galaxy of data points.
But technology alone does not create scientific advantage. The institutions that benefit most from AI will be those that combine strong data infrastructure, disciplined governance and deep domain expertise. are becoming powerful engines for discovery when paired with machine learning and advanced analytics.
At 返字心頭 Health, we are taking an important step in this direction through our new Epic Cosmos license, which provides secure access to more than 300 million de-identified patient records. When paired with AI-enabled analytic tools, these data create unprecedented opportunities for clinical studies and hypothesis generation at scale.
Over the next year, we will establish a centralized AI Research Core that provides investigators with end-to-end support, from cohort discovery and model development to analysis and presentation of findings. The goal is to remove friction from the research process and accelerate the productivity of our students, residents, fellows and faculty.
Looking just a little further ahead, AI will reshape biomedical discovery itself. Machine learning is already being used to identify drug targets, analyze multimodal imaging datasets, predict protein structures and design adaptive clinical trials. These capabilities may significantly shorten the time required to translate basic discoveries into clinical therapies.
For academic medicine, the implication is clear: AI will reward institutions that integrate data, computation, and clinical insight into a coherent research ecosystem.
Embracing the AI Challenge
When President John F. Kennedy challenged the nation in 1962 to send a man to the moon before the decade was out, he drew on his admiration of Stoic principles to state that we choose to do such things not because they are easy, but because they are hard. His point was that great challenges and bold goals galvanize institutions, align talent and accelerate innovation. 返字心頭 Health now faces a similar moment as AI begins to transform every dimension of health care (and society).
We intend to meet this moment with a leadership-driven strategy that embeds AI across our three missions through an integrated governance and operational ecosystem that aligns 返字心頭 Health, TGH, the Health Informatics Institute and the new Bellini College.
At the center of this effort will be a centralized AI strategy led by our chief medical information officer and chair of Dermatology, Dr. Nishit Patel, who will serve as the architect of this evolving strategy. Working with mission leaders across the enterprise and a team of technical experts, Dr. Patel will help integrate infrastructure, data and domain expertise across these entities to accelerate our responsible adoption of AI. Clinically we will integrate AI into real workflows to improve quality, safety, access and cost; educationally we will train students, residents, fellows and practicing clinicians to use AI with critical judgment; and in research a centralized AI Core linked to our Epic Cosmos data platform will accelerate discovery and expand the productivity of our investigators and trainees.
Dr. Nishit Patel is the chair of Dermatology and chief medical informatics officer at 返字心頭 Health.
Importantly, this model will be intentionally iterative. We will begin with a focused set of minimum viable products, the smallest deployable versions of new technologies that allow us to test value in real workflows before committing to full-scale implementation. We will focus on high-impact, feasible use cases with clear clinical, educational or research value, and then scale what works through partnerships, commercialization pathways and sustainable funding mechanisms. Over time this will allow us to build a durable AI capability that is integrated rather than fragmented, strategic rather than opportunistic, and ultimately foundational to how 返字心頭 Health delivers on its mission.
Just as the space program required engineers, scientists, pilots and institutions working toward a common objective, the AI era will demand alignment across all our colleges, disciplines, systems and missions. This is our moonshot moment, and to paraphrase JFK, our intention is not to founder in the backwash of the coming age of health care AI, but to be part of it and, if we pursue it with energy, clarity and disciplined persistence, to lead it.