Artificial intelligence is reshaping health IT at a pace that many healthcare organizations are struggling to match. From ambient documentation and clinical decision support to agentic AI that can act inside the EHR, the tools available to clinicians and administrators have multiplied rapidly and so have the questions about whether healthcare is ready for them.
Becker’s asked CMOs, CMIOs, CNEs, and other health IT leaders to identify the biggest threats in health IT today. Their answers converged on a familiar tension: the gap between the speed of innovation and the governance, clinical oversight, and workflow integration needed to deploy it safely.
Whether the concern was AI adopted without clinician input, models that drift unmonitored, or flawed processes scaled faster than they can be corrected, leaders across the country are sounding the same alarm technology must earn its place in patient care, not simply occupy it.
Mohamed Rami Nakeshbandi, MD. Vice President and CMO, Executive Director of Master of HealthCare Administration Program at Downstate Health Sciences University: I think the most concerning trend in health IT is the rapid adoption of AI and other technologies without adequate governance, validation, and evidence of effectiveness. Hospitals and leaders are being overwhelmed by AI and other technology vendors knocking on our doors every day, each promising to improve quality, efficiency, and revenue cycle. But we have very limited standardized ways to objectively evaluate which solutions truly meet our clinical need, deliver meaningful value, and have the real return on investment. The danger is that we end up adopting a technology because it is available and compelling rather than it is proven to improve efficient care. As health systems, we need to become better in separating innovation from implementation and promise from evidence.
David Murphy, MD, PhD. Vice Chair for Quality, Department of Medicine at University of Colorado School of Medicine (Aurora, Colo.): The most dangerous trend in health IT is allowing automation to advance faster than our ability to preserve meaningful human judgment and maintain safe, understandable healthcare systems. Individual clinicians and entire health systems are under enormous pressure to provide care faster and more efficiently, and AI and other technologies offer powerful tools to help achieve this. However, as these tools increasingly shape clinical decisions and care delivery at individual and population levels, we must ensure that efficiency does not come at the expense of quality and safety. This requires designing technologies and broader systems so clinicians can exercise meaningful judgment, failures remain visible, and emerging problems can be recognized and addressed.
Enitza George, MD. Associate Professor and Chief Population Health Officer at SUNY Downstate Health Sciences University (New York): The most dangerous trend in health IT is automating systems we haven’t examined, validated and redesigned. We’re putting increasingly powerful AI and digital tools on top of workflows that were built for a different era. Technology scales whatever system we give it; including its inefficiencies and bad assumptions. So, before asking whether a process can be automated, we need to ask whether that process should exist in its current form at all. Not evaluating the systems prior to automation is dangerous because once you automate a flawed process, you don’t just preserve the flaw, you scale it. You make it faster, harder to see, and potentially much harder to undo. In healthcare, the consequence isn’t just wasted money or inefficiency. It can affect access, equity, clinician behavior, and ultimately patient outcomes, and safety. So my concern isn’t that technology won’t transform healthcare. That will certainly happen, for good and for bad. My concern is that unless we are careful, we risk becoming extraordinarily efficient at doing the wrong things.
Deepa Velayadikot, MD. Medical Director of Care Coordination; Regional Medical Director, Hospital Medicine at Cooper University Hospital: The most dangerous trend in health IT is the growing expectation that AI is the solution to every problem. While AI has significant potential, many AI initiatives are being adopted because they are trendy rather than because they address a clearly defined clinical or operational need. This can lead to substantial investments with little measurable return on investment, increased clinician burden, poor workflow integration, potential safety and quality concerns, and a loss of trust when AI tools fail to deliver meaningful outcomes.
A more effective approach is to focus on “Relevant AI”: technologies that have been thoroughly evaluated by clinical, operational, and content experts and demonstrated measurable value. AI should solve a specific problem, improve care delivery, enhance efficiency, or support better decision-making, rather than being adopted simply because it is the latest trend.
Without thoughtful evaluation, strong governance, and alignment with organizational goals, AI risks becoming an expensive buzzword rather than a meaningful tool for healthcare transformation.
Salim Saiyed, MD. CMIO of UT Health Austin: The most dangerous trend in health IT is chasing the next shiny technology. Healthcare is deploying more technologies than ever, yet many clinicians feel less productive and more burdened. We are frequently adding more complexity without adding clinical value. The biggest danger is adding to cognitive overload — more alerts, more inbox , more documentation, etc. If we are not intentional, technology can actually amplify the problem by generating more information for clinicians to process rather than reducing the burden.
Terry McDonnell, DNP, RN, ACNP-BC. Senior Vice President and Chief Nurse Executive at Duke University Health System and Vice Dean for Clinical Affairs at Duke University School of Nursing (Durham, N.C.): The single most dangerous trend in healthcare IT is the development of solutions without our frontline nurses, physicians, and other key clinical stakeholders being left out of the equation. The best solutions are driven by the people who are closest to the work and we know from a change management and adoption perspective when the front-line stakeholders are involved then the solutions are embraced and care transforms. There are a lot of companies out there that are very well intended but not developing with a co-development viewpoint and having clinical advisors or an advisory board is different from co-development.
The second piece to this is when executives make decisions to purchase and implement solutions that have not had the patient and clinician involvement to ensure that the right solutions and changes are taking place. When the wrong solutions are brought forward that creates waste and I think we can all agree that we can’t afford to introduce waste into our already complex systems.
Anitra Williams, DNP. Area Chief Nurse Executive and Interim COO at Kaiser Permanente, Central Valley: In my opinion, the most dangerous trend in health IT is allowing technology to move faster than our ability to safely integrate it into clinical care. AI is a great example. The risk is not AI itself, rather it is deploying it without strong clinical governance, transparent validation, cybersecurity safeguards, and the voice of the clinicians who will ultimately use it. As a CNE and COO, I worry about technology that adds complexity instead of reducing it, creates more data without better decisions, or automates processes without understanding clinical consequences. The real opportunity is to use technology to give time back to caregivers, reduce cognitive and administrative burden, and improve the reliability and safety of care. Innovation should never be measured by how quickly we implement a new tool, but by whether it measurably improved the experience and outcomes of patients and the people caring for them.
Hetal Rupani, Senior Director, Business Intelligence and Analytics, Johns Hopkins Medicine (Baltimore): One of the most dangerous trends in health IT is the assumption that technology adoption equals transformation. It doesn’t.
The real risk is deploying tools faster than organizations can operationalize them — without the workforce training and upskilling, workflow redesign, governance, and accountability needed to turn technology into value. In health care, successful digital transformation is not measured by implementation alone; it is measured by whether technology improves care delivery, reduces burden on clinicians, and drives better patient outcomes.
A second risk is the blurred line between what EHR vendors enable and what health systems must own. EHR platforms are foundational, but the responsibility for making technology scalable, usable, and clinically meaningful sits with health systems. When that
distinction is not clear, organizations can end up with expensive systems that are technically live but operationally underperforming.
At the end of the day, the most dangerous trend is not innovation itself — it is innovation without adoption, alignment, and measurable impact.
Pradeep Kadambi, MD. Vice President and Chief Medical Officer of Upper Valley Medical Center (Troy, Ohio): There are several things that are ongoing concurrently. Obviously, cybersecurity should at the forefront of every organization. There are risks of autonomous Artificial intelligence (AI)-driven cyberattacks and an over-reliance on technology without adequate human oversight. In addition, there should be back-up plans as to how to continue to manage patients safely and keep operations running if that were to happen.
Another dangerous trend in health IT revolves around unregulated AI use. Recent reports indicating the misuse of consumer AI chatbots (like ChatGPT, Gemini, and Claude) by patients and clinicians for unvalidated medical guidance is rather alarming and ranks as the single most critical threat to patient safety today.
Hasan Ahmad, DO. Vice President and Associate CMIO of Parkview Health (Fort Wayne, Ind.): The most dangerous trend in health IT? We’re deploying AI faster than we’re learning how to govern it.
Too many organizations treat governance as a procurement checkbox — something you do once before go-live. We validate a predictive model, turn it on, and then let it run for years. Meanwhile the patient population shifts, documentation patterns change, and performance quietly drifts. Nobody’s watching.
The other danger is what we don’t turn off. Every health system has legacy alerts firing thousands of times a day that no longer earn the interruption. At Parkview Health, we’ve removed more than 100,000 interruptive alerts per month, and I’ll tell you — deimplementation is a clinical skill. Most organizations haven’t built that muscle yet. If we drop AI into environments already saturated with noise clinicians have learned to ignore, they’ll ignore the AI too.
The answer isn’t slowing down. It’s treating AI governance like a clinical discipline: continuous surveillance, clear thresholds for intervention, and the willingness to retire what isn’t working.
Reetu Singh, MD. Senior Vice President and Chief Medical officer of Saint Francis Health System (Sussex County, Del.): The most dangerous trend in health IT is the growing number of solutions built without clinicians at the table. Too many technologies are developed in search of a problem and marketed with promises of better outcomes before demonstrating real-world value. Healthcare innovation succeeds when clinicians, operators, and technologists work together to solve actual patient care challenges and deliver measurable results. Technology should solve real clinical problems and be validated in practice, not simply introduce new tools in search of a use case.
Tinu Tadese, MD. System CMIO of Boston Medical Center: There is no shortage of exciting technology that clinicians and operational leaders want to bring into our health systems. But implementation is only the beginning. The real work is aligning workflows, establishing ownership, and ensuring teams have the capacity to continuously optimize and sustain it. And we have to be careful not to define ROI only in financial terms—reducing clinician burden, improving care, and giving time back to our workforce are also meaningful returns.
Tomi Kolade, MBBS. Assistant CMIO at UTHealth Houston: The most dangerous trend in health IT is building agentic AI for clinicians without clinicians meaningfully involved in designing the agents.
We are moving from AI that summarizes and suggests to AI that can increasingly act: preparing orders, managing inbox work, coordinating care, and executing multi-step workflows inside the EHR. That shift changes the stakes entirely. A poorly designed screen frustrates a physician. A poorly designed agent acts on the wrong assumption: routing a message that needed a physician’s eyes, closing a loop that wasn’t actually closed, executing a workflow with no one watching until the outcome is already downstream.
As we enter the agentic era, physicians cannot simply be end users brought in at the end to validate what has already been built. They need to be embedded in the design and validation loop from day one, helping define what the agent should do, when it should act, when it should stop, and when human judgment must take over.
The future of the EHR should not be AI designed for clinicians. It should be AI designed with clinicians.
Stephen Bardoczi. Vice President, Planning and Post Acute Care Emeritus at Witham Health Services (Lebanon, Ind.): The most dangerous trend is failing to fully assess your organization’s true IT needs amidst the noise and hype from external sources, such as vendors and competitors. Additionally, organizations often struggle to truly understand the resources required to manage changes or implement new systems.
The organizations that succeed won’t be the ones moving fastest. They’ll be the ones disciplined enough to fix the foundation first, simplify where they can, and require every new capability to earn its place before it’s trusted with patient care.
Witham Health ServicesLewis W. Marshall Jr., Chief Medical Officer, NYC Health + Hospitals/Lincoln Hospital: Health IT is developing at a tremendous speed. New systems and programs are being developed on a daily basis. Since we all have some of the shiny object syndrome we need to be wary as we sort through the shiny objects. For clinicians, patients and institutions the speed which we develop understanding of these shiny objects and if they are accurate, have limited bias and can be trusted is much slower than the development process. The difference between the speed of development and adoption can be dangerous in that we may tend to adopt untested health IT because it is available. A second speed issue is that trust can develop fast. When an IT system is embedded in our EHR, we give it an aura of authenticity even if the model is uncertain. This may lead to rapid development in trusting the technology which can lead to medical errors directly affecting patients. We may need to speed up our understanding of these new systems to better control how we use them in practice.
Tony Sillemon, PsyD. Director, Community Health at Alta Bates Summit Medical Center (Oakland, Calif.): One of the most dangerous trends in health IT is allowing technology to advance faster than the systems, safeguards, and workflows needed to support it. AI, automation, and digital health tools have tremendous potential, but if they are implemented without enough attention to data quality, privacy, equity, clinical oversight, and the realities of frontline care, they can unintentionally widen disparities or create new risks for patients.
In my work with EBAC, which serves patients living with HIV and others who may face significant barriers to accessing care, I have seen firsthand how technology can either improve access or create another obstacle. Digital tools should make it easier for patients to connect with care, manage medications, communicate with their care teams, and remain engaged in treatment. But if systems are not designed around the needs of the populations we serve, particularly those who may experience digital, socioeconomic, language, or health literacy barriers, technology can actually deepen the very inequities we are trying to address.
The goal should not simply be to adopt more technology. It should be to make sure technology improves access, supports clinical decision-making, reduces administrative burden, protects patient trust and privacy, and ultimately leads to better outcomes for every patient.
James Matera, DO. Senior Vice President for Medical Affairs and CMO at CentraState Medical Center (Freehold, N.J.): I feel the most dangerous trend isn’t that AI will replace physicians. It’s that physicians may gradually surrender judgment to AI without realizing they’re doing it. AI can augment clinical judgment, but it cannot become a substitute for clinical accountability. The fear that this will also lead to clinical “Deskilling” has me concerned for the future stewards of healthcare.
The question for a CMO isn’t simply, ‘Does the AI work?’ It is more: ‘Who is accountable when it doesn’t, how will we know when it stops working, and can a clinician still intervene before the patient is harmed?’ These are key factors in High Reliability and Fair and Accountable Culture.
Vitaly Herasevich, MD, PhD. Professor of Anesthesiology and Medicine, Mayo Clinic (Rochester, Minn.): Blind believe that AI/LLM/EMR/Health IT etc. will solve all patients’ problems.
Shane Borkowsky, MD. Associate Chief Health Information Officer at UI Health (Chicago): Regarding the question of “what is the most dangerous trend in health IT?” I have a few thoughts. We are all excited for new AI tools- documentation, ordering, embedded CDS and guidance during the visit, and other tools. These tools can remove significant burden from providers, who currently spend more time on documentation (precharting, note completion, LOS selection) than they do in direct communication with the patient.
My concern with these tools is that they are only as powerful as the data they are working from. We all have inaccuracies in our EHRs- typos, missed dictations, copy/forward or copy/paste without updating or correcting errors. Taken individually, it may not seem big. But when an incorrect diagnosis or result or document goes unchecked and is now taken as fact, the resultant tools, including documentation, summaries, orders, patient education, may propagate the errors. We also know that AI is not perfect, and ambient dictation can be incorrect, confusing the roles of provider and patient, or family member in the room with either patient or provider.
When we rely on the document without reviewing the content, it is easy to miss small errors in attribution or in content. We are all stewards of our EHR, and we need to take the responsibility to ensure accuracy of the content, as should all users of any EHR.
Pooja Vyas, DO. System Vice President of Care Coordination Liaison, Physician & Provider Advisement at SSM Health (St. Louis): The most dangerous trend in health IT is the belief that technology alone can solve healthcare’s problems, particularly when it is implemented without redesigning workflows, engaging clinicians, or preserving clinical judgment. Too often, new tools add clicks, alerts, documentation requirements, and administrative work rather than reducing cognitive burden and allowing clinicians to focus on patients. At the same time, the growing reliance on algorithms and automation risks distancing clinicians from decisions that affect patient care, medical necessity, documentation, and reimbursement. Healthcare does not need more data or technology layered onto inefficient processes; it needs actionable intelligence and thoughtfully designed solutions that connect clinical care with operational and financial outcomes. Technology and AI should augment clinician expertise, not replace it, and should be implemented with clinicians rather than done to them.
Gary Zimmer, MD. CMO of Mohawk Valley Health System (Utica, N.Y.): AI growth and development is occurring faster than provider governance can keep up.
Also, trend toward single solution vendors creates “too big to fail” (i.e. Epic is building rev cycle, CRM and other back office functions to squeeze out other vendors.
De-Ann Pillers, MD, PhD. Dharmapuri Vidyasager Endowed Professor and Medical Director, Chicago of UIC NIDCAP Training Center, University of Illinois, Chicago: I would say that the most dangerous trend is to rely on MyChart as a communication tool for patients. Patients need a personal interface, not just try to understand notes or lab results. Most people are not medically sophisticated and can get hung up on things that are not of much consequence. It is getting harder and harder to accomplish that because doctors and extenders are booked weeks out and you can’t get an appointment. There is an art and a science to medicine, and the art is being given short shrift these days.
Curtis Cole, MD,. Vice President and Chief Global Information Officer at Cornell University (New York City): I think the most dangerous thing happening in healthcare IT is the abuse of IT systems by malignant government actors. When agencies like DOGE or ICE use medical records collected to care for people in order to pursue a political agenda against them it causes permanent damage to the doctor patient relationship and our overall trust in the healthcare system. Aggregating data in unified EHRs with interoperability is a huge victory for improving care. When that same data integration is used for political purposes like opposition to abortion, immigration, or LGBT care, it undermines the whole system which relies on a foundation of trust. I just moved my own care to a practice that doesn’t have Epic. Initially I was worried about this. Now I think it might be a feature.
Mahesha Amin, MD. CMO, Clinical Institutes at BayCare Health System (Clearwater, Calif.): I think unvetted generative AI and clinical chatbots.
Nipa R. Shah, MD. Chair of Community Health and Family Medicine at University of Florida, College of Medicine (Jacksonville): The most dangerous trend in health IT is the sheer pace of change. New technologies, particularly AI, are evolving so rapidly that many healthcare organizations struggle to make confident strategic decisions about whether to buy, borrow, build or get ‘bot’ solutions. By the time a decision is made, the landscape may have already shifted.
The result is often organizational paralysis. Leaders become hesitant to commit significant resources to a technology that could be outdated, replaced, or fundamentally transformed within months. While caution is understandable, the greater risk may be stagnation — delayed decisions, missed opportunities, and an inability to realize the benefits of innovation.
Organizations that learn to make thoughtful but timely decisions will be best positioned to improve care, reduce costs, and create value in an environment of constant change.
Elizabeth Johnson, PhD. Assistant Professor in the College of Nursing and Co-Director of the Biomedical Innovation for Research and Development Hub at Montana State University (Bozeman): The most dangerous trend in health IT is fragmented innovation. Too often, cohesion is missing across the product lifecycle, if co-design with end users and developers occurs at all through adoption, implementation, surveillance, and post-deployment monitoring.
A narrow, enterprise-by-enterprise approach creates blind spots where mistrust, misrepresentation, and misunderstanding can take root, with ripple effects on downstream ideas and decisions. At the same time, access to innovation remains shaped by capital availability, unstable startup partnerships, and traditional pathways that leave rural, frontier, and other under-engaged communities at the margins. Without a more inclusive and integrated approach, health IT may deepen disparities rather than improve care. We are all stewards of health technology and must approach that responsibility collectively, not competitively. We are not simply building products for “customers,” and profit cannot be the sole measure of value when long-term sustainability remains unresolved. We are building for ourselves, our families, and our neighbors and, ultimately, for a global future defined by possibility rather than which groups can shoulder burden.
April Kranz, DNP, RN. Vice President of Procedural Areas at Emory Healthcare (Atlanta): I believe the most dangerous trend in health IT is the rapid adoption of technology without equal attention to workflow integration, governance, and clinician usability. Artificial intelligence, clinical decision support, and increasing automation have tremendous potential to improve care, but poorly designed or poorly implemented systems can contribute to alert fatigue, documentation burden, communication failures, and overreliance on technology. Technology should enhance (not replace) clinical judgment, and every new solution should be evaluated based on its ability to improve outcomes, reduce cognitive burden, and support safe, reliable care.
James Solava, DO. Associate Chief Medical Informatics Officer, Northern Region, WVU Health System (Morgantown, WVa.): The most dangerous trend in health IT is the accelerating fragmentation of care, and AI risks speeding up a problem that was already underway. Continuity, seeing the same physician over years, is a clinical advantage, and not just for the reasons people assume.
AI can synthesize a decade of medical records faster than any human. But the chart isn’t the patient. A physician who has followed someone for years develops context that can be difficult to capture in data. Just as importantly, we often develop relationships with patients’ families and caregivers. When an older patient begins losing their memory, for example, the patient may not recognize that anything has changed. A spouse or adult child often does, and they bring that concern to the physician they know and trust.
AI can reconstruct the longitudinal record. It cannot reconstruct the longitudinal relationship.
As we use AI to make healthcare more accessible and efficient, we should be careful not to mistake information continuity for relationship continuity. The goal should be to use AI to strengthen the patient-physician relationship, not convince ourselves that technology has made that relationship less necessary.
Nariman Heshmati, MD. Chief Clinical Officer at Lee Physician Group & Ambulatory Care (Fort Myers, Fla.): The most dangerous trend in health IT is the rush to adopt artificial intelligence without adequate governance, validation, and clinical due diligence. I am increasingly seeing legacy solutions repackaged as AI—solutions that didn’t have efficacy or utility before. AI will be transformational in healthcare but labeling something as AI doesn’t make it the right solution.
Alexander Levit, MD. Medical Director- Hospital at Home for Lee Health System (Fort Myers, Fla.): While the situation has not settled, there is a real threat of a functionally low competition environment for EHR AI token pricing. While the frontier AI labs (OpenAI, Anthropic, etc.) are, as of August 2026, in a healthy competitive environment, Epic will have strong pricing control on the emerging agentic AI space inside enterprise environments. This is particularly disadvantageous to health systems that are too small to build the internal AI platforms of a Mayo Clinic, or to command the vendor terms of a Kaiser Permanente.
Toyosi Olutade, MD. Chief Medical Officer at UnityPoint Health-Quad Cities (Moline, Ill.): I think the most dangerous trend in health IT is allowing AI adoption to move faster than our ability to govern it. AI has tremendous potential to improve documentation, clinical decision-making, patient communication, and operations, but healthcare cannot treat it as just another technology rollout. These tools are increasingly influencing clinical care and the decisions clinicians make every day, so the stakes are much higher.
The greatest risk is that we scale error just as quickly as we scale value. If the underlying data are flawed, models drift, or clinicians begin to rely on technology at the expense of clinical judgment, the impact can spread rapidly across an entire health system. The opportunity is enormous, but clinical governance, data integrity, cybersecurity, and meaningful human oversight have to advance at the same pace as innovation.
Ilham Atir. Laboratory Director at Bergen New Bridge Medical Center (Paramus, N.J.): The most concerning trend in health IT right now is the rush to deploy autonomous AI agents before we have the institutional governance, clinical validation, and security oversight to match. Because these non-deterministic systems can alter their internal reasoning paths from one run to the next, rushing them into clinical workflows invites hidden biases, dangerous hallucinations, and catastrophic patient care errors. Preventing widespread harm requires healthcare organizations to enforce uncompromising guardrails and mandate a continuous human-in-the-loop requirement for every decision.
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
