Top AI Healthcare Voices in 2026: The People Actually Shaping AI in Medicine
Offcall’s 2026 Top AI Healthcare Leaders to Follow
Meet and Follow the People Driving AI in Medicine Forward
Artificial intelligence is everywhere in healthcare headlines these days.
It drafts notes. Flags sepsis. Screens imaging. Predicts deterioration.
But beneath the demos and funding rounds is a quieter question: Who is actually shaping how AI is used in medicine? Because AI in healthcare is not just a technical breakthrough. It is a clinical intervention. An ethical decision. A workflow redesign. A trust contract between patient and physician.
At Offcall, we are committed to lifting up the voices of physician and healthcare leaders on the frontlines, who are doing their part to shape AI for the better and put physicians at the center.By lifting up their voices, we are hoping to help expose others to their work and also to recognize their amazing contributions: helping to define guardrails, expose bias, integrate systems into real hospitals, and insist that medicine moves forward responsibly.
Without further ado, here are the top voices across medicine who are shaping AI for the better.Here are the voices doing that work.
1. Dr. Eric Topol — Restoring Humanity Through AI
Where some see AI as automation, Eric Topol sees restoration.
In Deep Medicine, Topol argues that artificial intelligence could give clinicians back what bureaucracy has taken: time, diagnostic depth, and meaningful patient connection. AI, in his view, is not about replacing physicians. It is about liberating them from administrative overload.
Through his Substack, Ground Truths, he curates and critiques emerging research, helping physicians distinguish validated breakthroughs from premature enthusiasm. His analysis is careful, evidence-driven, and grounded in patient outcomes.
In his influential Nature Medicine article, High-Performance Medicine, Topol outlines a model where human expertise and machine intelligence converge — not compete.
Topol’s work consistently reframes AI as a force capable of deepening the human side of care, not diluting it. Listen to his amazing podcast interview with Offcall’s co-founder Dr. Graham Walker here.
2. Dr. Shiv Rao — Reducing the Documentation Burden
As founder and CEO of Abridge, Rao has helped develop generative AI tools that transform clinician-patient conversations into structured clinical documentation – reducing the administrative burden that contributes to physician burnout (see: Abridge: AI for Clinical Documentation).
Rather than replacing clinical reasoning, these systems are designed to operate in the background, supporting documentation while allowing physicians to stay present with patients.
Rao has spoken publicly about the importance of building AI that earns clinician trust by prioritizing accuracy, transparency, and workflow integration, particularly in high-stakes medical environments (follow his updates: Shiv Rao on LinkedIn).
Healthcare depends on attention. Attention depends on cognitive bandwidth. And cognitive bandwidth depends on systems that remove unnecessary clerical load. Rao’s work positions generative AI not as a replacement for clinicians, but as infrastructure that restores their focus.
3. Dr. Bob Wachter — Chronicling Digital Medicine’s Cultural Shift
As the author of the New York Times bestseller The Digital Doctor: Hope, Hype, and Harm at the Dawn of Medicine’s Computer Age, Dr. Robert Wachter has long explored how technology reshapes clinical practice. He is widely recognized for his thoughtful examination of electronic health records, clinical decision support, and the evolving role of artificial intelligence in care delivery.
His 2026 book, A Giant Leap: How AI Is Transforming Healthcare and What That Means for Our Future, serves as a definitive roadmap for navigating the "Hemingway moment" of AI — where change happens gradually, then suddenly.
In numerous articles and commentary—including his contributions to NEJM Catalyst and his widely followed insights on UCSF Medical Grand Rounds — Wachter discusses how AI tools must be integrated into clinical workflows with a focus on safety, evaluation, and clinician trust. His perspective brings a historian’s depth and a clinician’s pragmatism to the conversation, reminding us that AI doesn't have to be perfect — it just has to be better than the status quo.
Healthcare depends on cultural adoption as much as technical performance. Cultural adoption depends on trust, usability, and evidence of benefit. And trust depends on leaders who analyze—not simply applaud—technological change.
Wachter’s work positions AI as a pivotal chapter in healthcare’s broader digital evolution, emphasizing that successful innovation must respect context, history, and human judgment.
Listen to How I Doctor podcast interview with Graham
4. Matt Sakumoto — Frontline Generative AI Literacy
Clinicians are already encountering generative AI in the tools they use every day, and Dr. Matt Sakumoto has been one of the primary voices helping frontline physicians navigate this transition. A virtual-first family physician, Sakumoto emphasizes that while large language models (LLMs) like ChatGPT offer transformative potential for reducing administrative burden, they must be evaluated through a rigorous lens of clinical safety and accuracy. In his widely followed LinkedIn commentary and appearances on podcasts like The Digital Patient and One Click At a Time, he outlines specific concerns around "hallucinations" in clinical summaries and the necessity of maintaining data privacy in digital workflows.
He is a vocal advocate for "Digital Empathy," arguing that while generative AI can help draft compassionate responses to patient portal messages, the final touch must remain human. Sakumoto’s 2024 and 2025 research, including his work in Telehealth and Medicine Today, focuses on creating guardrails—such as human-in-the-loop oversight and context-specific validation—to ensure that AI assistants support rather than supplant medical reasoning. He often uses the "Efficient Eddie" persona to educate peers on using AI tools to combat burnout without compromising care quality.
Sakumoto recently became the Chief Clinical Product Officer at Nabla, one of the leading AI ambient documentation companies. His work situates generative AI adoption inside that pragmatic framework, reminding the industry that even in a digital-first world, "the human touch" is the most critical component of the care team.Watch his AI webinar with Graham on Offcall here.
5. Spencer Dorn — Automating the Administrative Periphery
Healthcare AI succeeds only when it is applied to the right problems, and Dr. Spencer Dorn has been a leading voice in defining those boundaries.
As Vice Chair and Professor of Medicine at UNC, Dorn argues that medicine is fundamentally an "information processing discipline" that has become overwhelmed by its own data. Through his clinical practice and writing, he emphasizes that AI's most immediate value lies in summarization and administrative relief — reducing the "cognitive load" that leads to physician burnout. He cautions that while AI scribes are a significant first step, the true frontier is using AI to help doctors navigate the "overwhelming volume" of patient records and medical literature.
Healthcare depends on focus. Focus depends on reducing noise. And reducing noise depends on leaders like Dorn who view AI not as a replacement for clinical judgment, but as a tool to clear the path for it.
Listen to Spencer’s podcast interview with Graham here.
6. Michael Hobbs, MD — Bridging Clinical Practice and AI Innovation
A pediatrician and thought leader on healthcare technology, Dr. Michael Hobbs has shared firsthand observations of experimenting with AI health tools in primary care settings. He frequently notes how differences in clinical context—such as longitudinal patient relationships versus acute visits—can dramatically alter what an AI assistant should recommend and how it should be designed (see his viral reflection: "I tested AI health tools on my kidney stone").
His writing emphasizes that AI systems will behave differently depending on how they are trained and deployed, and that the human clinical context—continuity, history, and follow-up—matters just as much as algorithmic accuracy. In his role as a clinical co-designer at Elation Health, Hobbs highlights the distinction between pattern matching and contextual understanding. He argues that while models may be fluent in language, they can still miss the reasoning clinicians apply when integrating complex patient history into a differential diagnosis.
Such commentary underscores the need for AI systems to be developed with visibility into training priors and an awareness of longitudinal care rather than a "one-shot" emergency mindset. Healthcare depends on nuance in decision-making. Nuance depends on context. And context depends on tools that understand why clinicians consider certain data points as meaningful rather than treating them as isolated signals. Hobbs’ public reflections situate AI adoption within that real-world clinical reasoning framework—urging the design of systems that support context-aware care.
Watch Michael’s AI webinar with Graham here.
7. Dr. Geeta Nayyar — Translating Innovation Into Enterprise Strategy
A practicing physician and digital health leader, Dr. Geeta Nayyar has written and spoken widely about how AI and digital tools must align with clinician needs and patient outcomes, not just technical performance. In her work, she discusses how AI should be adopted responsibly, emphasizing training, equity, and clinician engagement so that predictive systems enhance rather than complicate care delivery (see her commentary: "AI Hopes & Fears in Healthcare").
Beyond commentary, Nayyar has shared insights on AI’s role in quality improvement and health system transformation in executive forums, focusing on frameworks for evaluation, governance, and human-centered design (follow her updates: Dr. Geeta Nayyar on LinkedIn). She is also the author of the bestselling book Dead Wrong: Diagnosing and Treating Healthcare's Misinformation Illness, which explores how technology and AI can either exacerbate or help solve the crisis of medical misinformation.
Healthcare depends on alignment between innovation and need. Alignment depends on leadership that bridges clinical insight and technology strategy. And that leadership depends on voices willing to translate potential into practice. Nayyar’s work places AI at that intersection—where innovation meets practical, equitable implementation.
Listen to Geeta’s How I Doctor podcast interview with Graham here.
8. Dr. Sarah Gebauer — De-Risking the Last Mile of Adoption
Before a tool reaches the patient, it must survive the scrutiny of clinical validation, a specialty of Dr. Sarah Gebauer.
Founder of Validara Health, Gebauer is an anesthesiologist focused on the governance and biosecurity of AI. She advocates for rigorous, pragmatic approaches to de-risk adoption, moving beyond "research demos" to production-grade engineering. Her work emphasizes that trust is earned through transparency, helping health systems navigate the complex regulatory and safety requirements of clinical AI.
Listen to Sarah’s AI webinar with Graham here.
9. Dr. Ethan Goh — Measuring the "Physician-as-Filter"
Healthcare AI is only as safe as the clinician who interprets it, a concept Dr. Ethan Goh defines as the "physician as filter."
As Executive Director of the ARISE Network, Goh leads large-scale evaluations of how physicians actually interact with LLMs. His research, published in venues like ResearchGate, highlights that while AI can outperform generalists on safety benchmarks, it still struggles with prioritizing clinically salient information under real-world time constraints. His work ensures that AI evaluation moves beyond static benchmarks to reflect the messy reality of the clinic.
Follow Ethan on LinkedIn here.
10. Dr. Jonathan Chen — Synthesizing Collective Clinical Wisdom
The true power of AI lies in its ability to synthesize the latent knowledge of the medical community, a vision championed by Dr. Jonathan Chen.
An Associate Professor at Stanford, Chen leads research into how machine learning can empower individuals with the collective experience of the many. He famously explored the "uncanny valley" of AI empathy in his award-winning essay "Who’s Training Whom?", noting that AI can sometimes navigate ethical dilemmas with more consistency than human clinicians. His work positions informatics not as a replacement for doctors, but as the only credible way to manage the escalating complexity of modern healthcare.
Follow Jonathan on LinkedIn here.
11. Dr. Jackie Gerhart — Re-Envisioning the Patient Experience
For AI to be actionable, it must live where care happens, a conviction held by Dr. Jackie Gerhart, CMO at Epic.
Gerhart advocates for deeply embedded AI that turns software into a "trusted colleague" rather than just a tool. Through her work at Epic, she has helped pioneer AI that reduces nurse documentation time by 85% and translates medical jargon into plain language for patients. Her mission is to use AI to restore the empathy and connection that defines the patient-provider relationship.
Listen to Jackie’s How I Doctor podcast interview with Graham here.
12. Dr. David Rhew — Bridging Global Platforms and Clinical Reality
As Global Chief Medical Officer and Vice President of Healthcare at Microsoft, Dr. David Rhew has written and spoken extensively on the transition of AI from a diagnostic novelty to a foundational "silent assistant" in clinical workflows. A physician-scientist and six-time U.S. patent holder, his work focuses on bridging the gap between cutting-edge technology and the practical needs of frontline clinicians (see: Dr. David Rhew on AI for Better Health).
Rhew is a leading advocate for the "Technology + Intervention = Outcome" formula, emphasizing that AI is only effective when it triggers a specific clinical action rather than just delivering data. He has been a driving force behind the Trustworthy & Responsible AI Network (TRAIN), a consortium aimed at operationalizing responsible AI principles through large-scale, diverse dataset testing and continuous monitoring for "algorithm drift."
In his interviews and keynote talks, such as on the How I Doctor podcast, Rhew discusses how agentic AI and ambient documentation (like the DAX Copilot) are essential for restoring physician attention and reducing burnout. He frequently highlights "oculomics"—the use of AI to detect systemic diseases through routine eye exams—as a prime example of how AI can scale preventative care and risk stratification.
Healthcare depends on accuracy and access. Access depends on efficiency. And efficiency depends on leaders who view AI as a tool to amplify human expertise rather than replace it. Rhew’s work positions AI as a "digital core" that empowers clinicians to practice at the top of their license while making healthcare more predictive and personalized.
Listen to David’s interview with Graham on How I Doctor here.
13. Dr. Robert Pearl — System-Level Reform and AI Economics
As former CEO of The Permanente Medical Group and author of Mistreated: Why We Think We’re Getting Good Health Care—And Why We’re Usually Wrong, Dr. Robert Pearl has written extensively about how technology must align with the economic and structural realities of healthcare. In his writing and keynote talks, he argues that predictive analytics and clinical decision support systems will only succeed when integrated into care delivery models that reward value and quality improvement rather than volume.
In his regular commentary for Forbes and on his official website, RobertPearlMD.com, Pearl addresses how AI should be deployed with clarity on return-on-investment, safety, and accountability. He emphasizes that innovation must align with the incentives that drive clinician behavior and organizational priorities. His perspective underscores that improving healthcare with AI is not just a technical problem—it is a policy and economics challenge as well.
In his more recent work, including his book Uncaring: How the Culture of Medicine Kills Doctors and Patients, he explores how AI can either bridge or widen the gap in the doctor-patient relationship. Healthcare depends on aligning incentives with patient outcomes. Incentives depend on clear measurement and governance. And governance depends on leaders who connect technology with systemic reform. Pearl’s work positions AI not as a standalone solution, but as a lever in a much larger effort to reshape healthcare delivery for better results.Follow Robert on LinkedIn here.
14. Dr. Ziad Obermeyer — Confronting Structural Bias
In Science, Obermeyer and his colleagues published Dissecting racial bias in an algorithm used to manage the health of populations, a study showing that a widely used healthcare algorithm systematically underestimated the health needs of Black patients (https://www.science.org/doi/10.1126/science.aax2342). The bias was not malicious. It was structural. But structural bias, when embedded in clinical decision tools, produces real harm.
The paper reshaped how hospitals, policymakers, and AI developers think about algorithmic fairness, highlighting that performance metrics alone do not guarantee equity. It forced the industry to confront how proxy variables and historical data can encode existing disparities.
Healthcare depends on trust. Trust depends on fairness. And fairness depends on systems that are audited, transparent, and accountable.
Obermeyer’s work makes clear that in medicine, algorithmic neutrality is not assumed. It must be proven.
15. Dr. Suchi Saria — Turning Prediction Into Prevention
In Nature Medicine, Saria and her team published research on a machine learning–based early warning system designed to detect sepsis earlier than traditional clinical methods by analyzing complex electronic health record data (https://www.nature.com/articles/s41591-018-0213-5). The study showed that validated predictive models could identify deterioration sooner and give clinicians critical time to intervene.
Beyond individual models, Saria has also written about the responsible application of machine learning in medicine, helping clinicians distinguish measurable progress from premature hype. Her work emphasizes validation in real clinical environments, operational integration, and outcomes that can be audited.
Healthcare depends on timely intervention. Timely intervention depends on reliable prediction. And reliable prediction depends on rigorous validation.
Saria’s work keeps AI accountable to that standard.Follow Suchi on LinkedIn here.
16. Dr. John Halamka — Governing AI at Institutional Scale
As President of the Mayo Clinic Platform, Dr. John Halamka has written and spoken extensively about responsibly applying AI in healthcare, from ambient documentation to predictive models that support clinicians while avoiding harm (see: Mayo Clinic’s John Halamka: We have to use AI responsibly). He advocates for an approach he describes as “think big, start small, move fast,” where tools are evaluated rigorously before scaling into practice.
Halamka has also co-authored Transforming Healthcare with AI, a book exploring how data science and AI are reshaping care delivery, from early disease detection to at-home medicine. He emphasizes that adoption must prioritize ethics, validation, and clinical accountability.
In his writing for the Mayo Clinic Platform Blog, he regularly discusses the delicate balance between innovation and patient safety, making the case that AI must be deployed with defined guardrails and continuous evaluation to maintain trust. Healthcare depends on tools that augment clinician judgment. Clinician confidence depends on evidence and oversight. And oversight depends on leaders willing to articulate how AI can be adopted safely. Halamka’s work situates healthcare AI inside that framework of responsible innovation.
17. Dr. Daniel Yang — The Principles of Responsible Scale
At the scale of a system like Kaiser Permanente, AI requires a rigorous ethical framework, a mission led by Dr. Daniel Yang.
As VP of AI and Emerging Technologies, Yang has outlined seven principles for responsible AI that prioritize transparency, equity, and clinician oversight. He champions tools that deliver a "return on health," such as the Advance Alert Monitor which saves 500 lives annually. His work ensures that AI serves as a tool for equitable, evidence-based care rather than a source of hidden bias.
Follow Daniel on LinkedIn here.
18. Brendan Keeler — The Architect of Interoperable Infrastructure
AI is only as good as the data it can access, and Brendan Keeler is the leading voice on the "plumbing" that makes this possible.
Known as the "Health API Guy," Keeler (a non-physician expert) demystifies the technical standards like HL7 and FHIR that allow AI to communicate with EHRs. He argues that healthcare standards are a feature, not a bug, and his work focuses on building the digital rails necessary to replace outdated workflows like the fax machine with modern, AI-ready data exchange.
Follow Brendan on LinkedIn here.
19. Dr. Adam Rodman — The Historian of AI Reasoning
To understand where AI is taking medicine, one must understand the history of medical logic—a perspective uniquely provided by Dr. Adam Rodman.
A general internist and historian of medical epistemology, Rodman uses his work and podcast, Bedside Rounds, to contextualize AI within the long arc of clinical reasoning. In his seminal essay "Can AI Make Medicine More Human?", he argues that we are shifting from "human-in-the-loop" to "human-on-the-loop," where AI performs the heavy lifting of data synthesis while humans provide the supervisory oversight and empathy.
20. Dr. Mark Sendak — Getting AI Out of the Ivory Tower
Successful AI integration requires a bridge between data science and clinical operations, a bridge built by Dr. Mark Sendak at Duke.
As the Co-Lead of the Health AI Partnership, Sendak has helped lead the responsible implementation of over 20 AI technologies for clinical care. His work focuses on creating the "roads, onramps, and bridges" needed for equitable AI adoption across all health systems, particularly in low-resource settings, ensuring that innovation benefits every patient, not just those at elite institutions.
21. Dr. John Lee — Informatics as a Synergy for Cognition
Informatics is the circulatory system of modern healthcare, and Dr. John Lee ensures it remains focused on enhancing human capability.
A Chief Medical Information Officer (CMIO), Lee’s philosophy centers on synergy between AI and human cognition. Through his work with AMIA, he emphasizes that AI's primary strength is the rapid pattern analysis of large datasets that humans simply cannot extract. He argues that routine adoption of AI will only occur when the technology is integrated directly into the clinician's workflow without increasing friction.
22. Dr. Nigam Shah — Predictive Modeling With Clinical Consequence
As Professor of Medicine, Co-director of the Stanford Center for Artificial Intelligence in Medicine & Imaging (AIMI), and Chief Data Scientist for Stanford Health Care, Shah’s research group develops machine learning techniques that analyze diverse real-world health data—including electronic health records, claims, and wearables. These insights build predictive models that clinicians use to improve care delivery and decision-making (see: Nigam Shah’s Stanford Profile).
A notable contribution from Shah and collaborators is the article "Developing a delivery science for artificial intelligence in healthcare," which highlights how AI must be evaluated not just for accuracy but for safe and effective clinical integration. Shah has also been involved in creating tools like MedHELM, designed to evaluate large language models on tasks that matter in healthcare, encouraging more relevant performance assessments (see: Evaluating AI on Healthcare Tasks).
Beyond research, he teaches AI in healthcare courses and co-founded Atropos Health to support evidence-based, real-world AI adoption. Healthcare depends on predictive insight and practical deployment. Practical deployment depends on models that are validated, context-aware, and integrated. And integrated AI depends on leaders who bridge research and clinical reality. Shah’s work anchors AI in the demands of real-world medicine.
Follow Nigam on LinkedIn here.
23. Dr. Bertalan Meskó — Preparing Physicians for Transformation
Known as The Medical Futurist, Meskó has written and spoken extensively about the real-world implications of AI and digital health for medical practice, helping physicians translate emerging technologies into everyday clinical insight. In his book The Guide to the Future of Medicine, he explores how artificial intelligence can support diagnosis, personalized care, and clinical workflows when it is integrated thoughtfully and ethically.
Through The Medical Futurist website and his active commentary on X (formerly Twitter), Meskó curates and interprets AI research, industry trends, and implementation case studies that matter to clinicians rather than technologists. His work reinforces that doctors need more than theoretical knowledge to engage safely with AI.
His commentary highlights not just what AI can do, but what it should do in healthcare contexts—emphasizing clinician agency, patient benefit, and culturally grounded adoption. Healthcare depends on informed use. Informed use depends on education. And education depends on voices willing to translate complexity into clinical relevance. Meskó’s work positions AI not as an abstract innovation, but as a tool that clinicians must understand, evaluate, and shape.
Follow Bertalan on LinkedIn here.
24. Dr. Harvey Castro — Generative AI in Real Clinical Settings
Dr. Castro has authored ChatGPT and Healthcare: The Key to the New Future of Medicine, a book that explores how large language models can assist clinicians with patient education, documentation support, and communication. He emphasizes that tools like ChatGPT must be used responsibly to be effective.
In articles such as "How to regulate generative AI in health care," Castro discusses both the promise and the regulatory challenges of generative AI systems in clinical environments, arguing that innovation must be balanced with safety and clinician oversight.
He regularly shares insights on the ethical application of generative AI—including concerns about hallucination, privacy, and the necessity of clinician verification. His work reinforces that AI should assist rather than replace professional judgment.
Healthcare depends on clinician trust. Clinician trust depends on responsible use. And responsible use depends on leaders who illuminate both potential and risk. Castro’s work positions AI exploration within a framework of practical skepticism and real-world applicability.
Follow Harvey on LinkedIn here.
25. Dr. Rachel Draelos — Bringing Interpretability to Medical Imaging
A radiologist and AI researcher, Draelos has written extensively about deep learning in medical imaging, including her work on Machine Learning for Medical Imaging, which helps clinicians and researchers understand how neural networks function in diagnostic contexts and how to evaluate their limitations. Her work emphasizes not just model performance, but interpretability and transparency.
Through her academic publications and technical writing on Glass Box Medicine, Draelos has examined how imaging models can fail, overfit, or pick up spurious correlations in medical datasets. She reinforces that clinical deployment requires careful validation beyond headline accuracy metrics (see her research via Duke Scholars).
Healthcare depends on accurate interpretation. Accurate interpretation depends on transparency. And transparency depends on researchers willing to scrutinize their own models. Draelos’ work positions medical AI not as a "black box" to be trusted blindly, but as a system that must be understood before it is adopted.
Follow Rachel on LinkedIn here.
26. Dr. Anthony Chang — Building AI Literacy in Medicine
As the founder of AIMed, Dr. Anthony Chang has written and spoken extensively about artificial intelligence in medicine, with a focus on how AI tools can support clinical decision-making without undermining physician judgment. In his book Intelligence-Based Medicine (and his co-edited work AI in Healthcare), he brings together expert perspectives on the current state of medical AI, clinical validation, and ethical considerations. His work helps clinicians differentiate validated innovation from premature hype.
Through his keynote talks and articles featured on the AIMed website and LinkedIn, Chang regularly translates complex AI concepts into clinically relevant terms. He emphasizes the need for education, safety frameworks, and multi-stakeholder collaboration in deploying AI tools in healthcare settings.
Healthcare depends on informed adoption. Informed adoption depends on clinician literacy. And clinician literacy depends on leaders willing to build bridges between machine learning research and everyday clinical practice. Chang’s work positions AI not as a technical novelty, but as a discipline clinicians must understand, evaluate, and shape.
Follow Anthony on LinkedIn here.
27. Dr. Sara Murray — Building the Infrastructure Behind Ambient AI
For AI to move from experimental pilot to essential clinical utility, it requires a foundation of rigorous governance and seamless integration, and Dr. Sara Murray has dedicated her career to constructing that very framework. As the Vice President and Chief Health AI Officer for UCSF Health, Murray oversees the strategic vision for AI across one of the nation’s leading health systems, having also served as the Associate Chief Medical Information Officer for Inpatient Care.
In her landmark 2025 research on the "Impact Monitoring Platform for AI in Clinical Care (IMPACC)," Murray outlines how a health system can move beyond simple implementation to continuous oversight. She argues that the shift toward "Ambient AI"—technologies like AI scribes that listen and document in real-time—must be supported by a robust backend that monitors for drift, bias, and clinical accuracy. Her work provides a blueprint for how institutions can evaluate both commercial tools and in-house models for "trustworthiness" before they ever reach a patient’s bedside.
Through her leadership in the Division of Clinical Informatics and Digital Transformation (DoC-IT), Murray emphasizes that AI's greatest value lies in its ability to reduce the "cognitive tax" on clinicians. In her recent discussions on the "Generative AI Era," she articulates that while large language models (LLMs) offer unprecedented speed, their true success is measured by how they enhance human connection—turning complex data into actionable insights that allow doctors to spend more time with patients and less time with screens.
Innovation must be built to last. Sustainability depends on infrastructure. And infrastructure depends on leaders who prioritize safety and equity in the digital architecture of care. Murray’s work positions ambient and predictive tools not just as new software, but as a fundamental redesign of the hospital environment, ensuring that the technology powering modern medicine remains invisible, reliable, and profoundly ethical.
28. Dr. Eric Poon — Operationalizing Digital Transformation
As Chief Health Information Officer at Duke Health, Dr. Eric Poon has written and spoken extensively about the real-world challenges and opportunities of AI adoption in health systems. He emphasizes that lessons from electronic health record (EHR) implementation should guide how AI tools are evaluated, governed, and scaled.
In his article "Déjà Vu? How Might Lessons Learned from Electronic Health Record Implementation Apply to Artificial Intelligence?", he and his co-authors argue that respect for human factors, strong governance, workforce readiness, and long-term strategy are essential for AI to improve care delivery rather than disrupt it.
Poon has also co-authored "Adoption of artificial intelligence in healthcare: survey of health system priorities, successes, and challenges," an analysis of how U.S. health systems are deploying AI tools and where operational barriers—such as immature technologies, financial constraints, and regulatory uncertainty—limit impact.
Through interviews and commentary, such as his discussion with HealthTech Magazine on balancing AI adoption with governance, Poon emphasizes the need for iterative evaluation, clinician engagement, and "just-in-time" learning as healthcare organizations introduce AI into clinical workflows.
Healthcare depends on innovation that improves outcomes. Innovation depends on structured learning. And structured learning depends on leaders willing to translate research into operational practice. Poon’s work places AI adoption inside that framework of evidence, governance, and real-world improvement.
29. Dr. Daphne Koller — Expanding AI Into Drug Discovery
As a co-founder of Coursera and later founder of the AI-driven drug discovery company insitro, Daphne Koller has brought machine learning to domains directly relevant to healthcare. Through insitro, she applies large-scale machine learning and biological data integration to accelerate drug discovery and understand disease mechanisms—work that intersects with healthcare AI by transforming early-stage research and therapeutic development (see: insitro’s Platform & Pipeline).
Koller’s interpretation of AI extends beyond images and text to high-dimensional biological data, emphasizing that models must be grounded in empirical validation and biological plausibility. Her leadership in this space has been featured in interviews and profiles that explore how computational models can change how we identify drug targets and predict treatment responses, such as her 2024 recognition as one of TIME's 100 Most Influential People in AI and her foundational STAT News interview on AI's role in medicine.
Healthcare depends on innovation at every stage—from diagnosis to therapeutics. Innovation depends on the ability to analyze complex data. And complex data analysis depends on leaders willing to tie machine learning to real, measurable scientific progress. Koller’s work positions AI not just as a clinical decision tool, but as a driver of discovery that can expand what is possible within medicine.
Follow Daphne on LinkedIn here.
30. Dr. Isaac Kohane — Foundational Biomedical Informatics
As the inaugural Chair of the Department of Biomedical Informatics at Harvard Medical School, Dr. Isaac Kohane has written and spoken extensively about how complex clinical and genomic data can be structured, understood, and used responsibly. His research emphasizes the importance of rigorous data governance and clinical context, illustrating how the integration of heterogeneous datasets is the only way to build machine learning that is both accurate and clinically meaningful.
Kohane has published extensively on using computational methods to uncover disease mechanisms, particularly through pediatric genomics and rare disease research. He is also the lead author of the influential book The AI Revolution in Medicine: GPT-4 and Beyond, which explores how generative AI will reshape the patient-clinician relationship.
As the Editor-in-Chief of NEJM AI, he leads the global conversation on the evidence required to move AI from research to the bedside. In his public commentary and interviews, he often highlights that AI systems trained on poorly governed data risk amplifying health disparities.
Healthcare depends on data that is accurate, accessible, and equitable. Equitable data depends on governance and context. And governance depends on leaders who build the infrastructure that supports trustworthy AI. Kohane’s work positions data stewardship as a prerequisite for meaningful AI in medicine.
Follow Isaac on LinkedIn here.
31. Daniel Nadler, PhD — Architecting Enterprise Data Systems
As co-founder and CEO of OpenEvidence, Nadler has built an AI-powered medical search and synthesis engine designed specifically for physicians. The platform helps clinicians navigate the exponential growth of peer-reviewed research by delivering evidence-based answers in real time. OpenEvidence uses specialized models trained on indexed clinical literature to provide precise, cited information at the point of care (see: About OpenEvidence).
In his interview on the No Priors podcast, Nadler explained his philosophy of treating doctors as "knowledge workers" and consumers. By focusing on reducing information overload, OpenEvidence has achieved rapid, organic adoption—now used by hundreds of thousands of verified U.S. physicians to solve complex patient cases.
OpenEvidence’s impact is further validated by its role as an official AI partner for major medical publishers, including the NEJM Group and the JAMA Network. These partnerships ensure that the platform’s AI-driven synthesis is grounded in the world’s most trusted medical evidence, supporting safer practice without replacing human judgment.
Healthcare depends on reliable, evidence-based insight. Evidence-based insight depends on tools that make complex data accessible. And accessible tools depend on leaders willing to build them with clinicians in mind. Nadler’s work positions AI as a practical "co-pilot" for physicians navigating the modern information landscape.
Follow Daniel on LinkedIn here.
32. Dr. Jennifer Joe — Designing AI Around Clinical Workflow
As a physician-entrepreneur and a prominent voice in digital health, Dr. Jennifer Joe has spent over a decade bridging the gap between medical innovation and clinical practice. As the founder of Medstro and MedTechBoston, she has been instrumental in creating ecosystems where physicians, technologists, and industry leaders can collaborate on the responsible adoption of new tools.
Currently serving as a Senior Director of Global Medical Strategy and Population Health at AstraZeneca, her work focuses on the intersection of large-scale medical data, emerging technologies like blockchain and AI, and their practical impact on patient outcomes (see: Jennifer Joe on ConV2X Blockchain & Tech).
Joe argues that for AI to succeed in high-stakes environments, it must be embedded directly into Medical Affairs workflows—supporting evidence generation, scientific engagement, and real-world data interrogation. She emphasizes moving beyond "innovation theater" to an operational reality where digital tools are evaluated for their ability to combat professional isolation and drive the adoption of more intuitive, clinician-friendly technology.
Through her leadership roles at the American College of Physicians (ACP) and the Massachusetts Medical Society, she has spearheaded initiatives to improve digital literacy among medical students and early-career physicians. Her advocacy for the "physician influencer" role, as discussed on The HealthXL Digital Health Podcast, encourages clinicians to take an active, vocal role in shaping the tools they use.
Healthcare depends on authentic engagement. Engagement depends on a sense of community. And community depends on leaders who provide the platforms for clinicians to lead the technological conversation. Dr. Joe’s work positions the physician not as a passive user of technology, but as its primary architect and ethical guide.
Follow Jennifer on LinkedIn here.
33. Dr. Keith Dreyer — Defining Standards in Imaging AI
A radiologist and computer scientist, Dr. Keith Dreyer has dedicated his career to the rigorous translation of artificial intelligence from experimental research into clinical reality. As the Chief Data Science Officer at Mass General Brigham and Chief Science Officer for the American College of Radiology (ACR) Data Science Institute, he oversees the largest portfolio of clinical AI deployments in the world, ensuring these tools are not only accurate but safe for patient care (see: Mass General Brigham AI Leadership).
Dreyer argues that the true challenge of healthcare AI is not just building algorithms, but managing their entire lifecycle. He has pioneered frameworks for "algorithm monitoring," emphasizing that even FDA-cleared models require continuous evaluation to detect "drift"—a decline in performance that occurs as clinical data and patient populations evolve over time (see his TIME 100 AI Profile).
Through his leadership at the ACR Data Science Institute, Dreyer has been a primary architect of standardized AI use cases and the ARCH-AI program, the first national quality assurance initiative for medical AI. His work focuses on integrating AI into existing physician workflows—such as automating complex image segmentation or triaging urgent findings—to solve the efficiency crisis caused by the explosion of medical data.
Healthcare depends on reproducibility. Reproducibility depends on validation. And validation depends on leaders who bridge the gap between high-level data science and frontline clinical necessity. Dreyer’s work ensures that as AI becomes increasingly autonomous, it remains anchored in transparency, ethics, and measurable patient benefit.
Follow Keith on LinkedIn here.
34. Dr. Monica Agrawal — Driving Clinical Implementation
An assistant professor of biomedical engineering and computer science at Duke University, Dr. Monica Agrawal focuses on how large language models (LLMs) and machine learning systems can be rigorously evaluated for clinical use. Her research documents the "evaluation illusion" of LLMs in medicine, urging context-aware assessments that capture real-world efficacy rather than relying on static, idealized benchmarks.
Agrawal has pioneered methods to bridge the gap between unstructured clinical text—the doctors’ notes, narratives, and jargon-heavy dialects found in electronic health records—and the actionable data required for decision support. In her work on few-shot clinical information extraction, she demonstrates how LLMs can be adapted to navigate messy real-world data with minimal labeled examples.
In collaborative studies analyzing how people seek health information, her team found that LLMs often struggle with real clinician–patient dialogue because they are trained on idealized, exam-style prompts. Her recent analysis, “What’s Up, Doc?”, reveals how models falter when faced with the ambiguity and false assumptions inherent in everyday patient inquiries.
Healthcare depends on evaluation frameworks that reflect real clinical contexts. Evaluation depends on recognizing where models succeed and where they fail. And recognizing failure depends on researchers willing to scrutinize how AI behaves outside academic benchmarks. Agrawal’s work positions clinical AI within that framework of rigorous, context-aware evaluation required for safe adoption.
Follow Monica on LinkedIn here.
35. Dr. Alice Chiao — Integrating AI Across Physician Networks
A former emergency medicine instructor at Stanford, Dr. Alice Chiao has moved into the emerging frontier of training artificial intelligence systems to respond more like real clinicians. Using reinforcement learning from expert feedback, she teaches models how to think, diagnose, and communicate in medically appropriate ways. Her insights are featured in the recent 2026 report "This doctor is training AI to do her job, and it’s a booming business," which explores the critical role of physician-led data labeling.
That reporting highlights how Chiao uses real clinical scenarios from her emergency department experience to guide AI responses. She argues that these systems should serve as tools that reduce administrative burdens—such as documentation or information retrieval—rather than replace human judgment. By acting as a "human-in-the-loop," she ensures the model's logic aligns with safe clinical practice.
In her public commentary on LinkedIn, Chiao emphasizes that AI should be taught in ways that preserve the traditional clinician's approach to care. She maintains that physicians must remain deeply involved in the adaptation of AI outputs before they ever influence patient decisions. Her work is a testament to the fact that "clinical intuition" can be codified, but only by those who have practiced it.
Healthcare depends on tools that think more like clinicians. Tools that think more like clinicians depend on clinician feedback. And clinician feedback depends on leaders willing to invest their expertise into AI training. Chiao’s work positions AI not as a black box, but as a system shaped by active clinician involvement.
Follow Alice on LinkedIn here.
36. Dr. Peter Lee — Generative AI Strategy at Enterprise Scale
As Corporate Vice President of Research & Incubations at Microsoft, Dr. Peter Lee has been a key figure in bridging advanced machine learning research with healthcare applications at scale. He has publicly articulated how artificial intelligence can be responsibly applied to healthcare challenges, emphasizing a balance between rapid innovation and patient safety.
Lee is a primary architect of Microsoft’s "AI for Good" initiative and a leading voice on the clinical utility of large language models. He is the co-author of the seminal 2023 book The AI Revolution in Medicine: GPT-4 and Beyond, where he discusses how tools like predictive analytics and natural language processing can assist clinicians with disease prediction and clinical summarization—provided they are grounded in ethical principles.
Lee also stewards Microsoft’s Responsible AI resources, which explicitly address healthcare-specific concerns such as bias mitigation and safety monitoring. In his interviews and industry keynotes, he highlights high-impact partnerships, such as the Trustworthy & Responsible AI Network (TRAIN), which aims to accelerate clinical research without compromising privacy or validity.
For clinicians and health system leaders navigating the "Hemingway moment" of AI—where change happens gradually, then suddenly—Lee’s commentary offers a research-informed vision. His work positions enterprise-level platforms as essential infrastructure designed to support—rather than supplant—clinical expertise.
Follow Peter on LinkedIn here.
37. Dr. Toby Cosgrove — Institutional Innovation Leadership
Dr. Toby Cosgrove brought a clinician’s perspective to strategic innovation long before “AI in healthcare” became a headline. As the former CEO and President of the Cleveland Clinic, Cosgrove oversaw major organizational efforts to integrate digital technology and data analytics into care delivery, helping one of the nation’s largest health systems embrace information-driven care long before many of its peers.
In his book Innovation: The Cleveland Clinic Way, Cosgrove examines how healthcare organizations can build cultures that systematically support innovation, including technology governance and multidisciplinary collaboration. His writing emphasizes that sustained improvement—whether through AI, decision support, or data science—depends on organizational maturity and strategic alignment rather than transient hype.
Now serving as an Executive Advisor for Google Cloud Healthcare and Life Sciences, Cosgrove continues to reflect on how technology should serve clinical priorities—improving access, reducing harm, and enhancing patient experience. In recent interviews, he has discussed how generative AI can alleviate physician burnout by automating administrative burdens, allowing doctors to return to the "joy of medicine."
Healthcare depends on nuance in decision-making. Nuance depends on context. And context depends on leaders who understand that technology must adapt to the clinician, not the other way around. Cosgrove’s work offers a model of how to integrate emerging technologies—including AI—into care delivery without sacrificing clinical values.
38. Dr. Karen DeSalvo — Public Health Governance in the Age of AI
Dr. Karen DeSalvo has shaped how healthcare systems and policymakers view digital transformation, from health equity to the ethical deployment of artificial intelligence. As the former National Coordinator for Health IT and current Chief Health Officer at Google, she has written and spoken extensively about the intersection of health data, digital tools, and population health. She notes that technology, including AI, must be governed carefully to improve outcomes without exacerbating disparities (see: Dr. Karen DeSalvo on Health Equity and Tech).
DeSalvo argues that digital health tools must be evaluated through lenses of equity, safety, and public health impact. In her contributions to Health Affairs, she addresses the use of predictive analytics and machine learning in public health reporting. She contends that investments in data infrastructure and responsible analytics can help identify community health needs earlier and target interventions more effectively. These discussions are central to AI’s promise in risk stratification and addressing structural inequities.
Beyond policy writing, DeSalvo leads Google’s efforts to ensure that AI models are helpful and inclusive, participating in global forums to articulate how governance frameworks—including standards for data interoperability—must evolve alongside AI capabilities. Her work emphasizes that for innovation to benefit all communities, it must be built on a foundation of ethical data use and high-quality evidence.
Healthcare depends on robust data and equitable design. Equitable design depends on governance and foresight. And governance depends on leaders who shape policy with both innovation and societal well-being in mind. DeSalvo’s work positions AI within a broader context of responsible digital transformation—where clinical technology serves health equity and the public good.
Follow Karen on LinkedIn here.
39. Dr. Amy Abernethy — Regulatory Science and Data Transparency
Innovation without regulatory clarity stalls. Dr. Amy Abernethy has shaped how data science intersects with regulatory oversight, particularly during her tenure as Principal Deputy Commissioner at the U.S. Food and Drug Administration (FDA). In her landmark FDA address, "Advancing Real-World Evidence and Artificial Intelligence," she outlined how real-world data (RWD) and AI can support regulatory decision-making only when paired with rigorous validation and patient safeguards.
Abernethy has co-authored seminal work in The New England Journal of Medicine examining how RWD and algorithmic tools must meet transparent scientific standards before they can influence care delivery. She argues that the transition from traditional clinical trials to a more fluid, data-driven approach requires a "learning health system" architecture that balances speed with safety.
Now serving as a co-founder and CEO of Highlandpoint Care Way and a board member for major healthcare entities, she continues to advocate for standardized data infrastructure to support post-market surveillance and drug development. Her work reinforces that AI-driven systems must earn trust through reproducibility, transparency, and measurable benefit.
Healthcare AI cannot bypass regulatory scrutiny; it must prove itself through the same rigorous evidence-based frameworks that govern traditional medicine. Abernethy’s work positions legitimacy and data integrity as the non-negotiable foundations of medical innovation.
40. Dr. Jesse Ehrenfeld — Physician Leadership in AI Governance
An anesthesiologist, informatician, and former President of the American Medical Association (AMA), Dr. Jesse Ehrenfeld has been a primary architect of physician-led governance for digital health. He has written and spoken extensively about the need for clinicians to define how AI tools are evaluated, implemented, and monitored, ensuring that technology serves the patient-physician relationship rather than disrupting it.
In his landmark AMA policy statements, Ehrenfeld spearheaded the adoption of ethical standards for AI in medicine. He advocates for "Augmented Intelligence"—a term he uses to emphasize that AI should enhance, not replace, human expertise—and insists on rigorous standards for transparency, clinician oversight, and equity. He specifically recommends that predictive algorithms be evaluated for bias and real-world impact before being deployed in patient care.
Through his contributions to the Journal of Medical Systems and his frequent commentary on LinkedIn, Ehrenfeld discusses how AI must be integrated into quality measurement and safety frameworks. He reinforces that physician input is essential for creating "clinically validated" AI that supports safer outcomes. His leadership during the 2023-2024 AMA term solidified the AMA's principles for AI, which now serve as a cornerstone for national healthcare technology policy.
Healthcare innovation requires clinician stewardship. Stewardship requires governance. And governance requires practitioners willing to shape standards instead of yielding to technology. Ehrenfeld’s work situates AI adoption within the broader context of ethical clinical leadership and professional accountability.
Follow Jesse on LinkedIn here.
41. Dr. Andrew Mellin — Coordinating AI Within Health Systems
As Vice President and Chief Medical Information Officer at Surescripts, Dr. Andrew Mellin has become a leading voice on how technology must support clinicians without becoming a hindrance. In a 2026 feature for Healthcare IT News, he predicted that the successful adoption of AI hinges on balancing optimism with realism. He emphasizes that provider organizations must apply rigorous scrutiny and governance to ensure these tools truly improve patient care rather than adding to a "digital tax" on clinicians.
Mellin is a vocal advocate for using AI-powered automation to tackle the healthcare industry's most persistent administrative burdens. In his 2026 healthcare predictions, he details how intelligent interoperability—including automated prior authorizations and real-time prescription benefit tools—can give clinicians back precious time with their patients. His commitment to human-centric design ensures that technology serves the clinical mission rather than forcing the clinician to serve the tool.
In his ongoing commentary, Mellin repeatedly notes that trust in AI is earned through responsible use, integration with established governance frameworks, and measurable impact on workflows. He argues that "invisible AI"—technology that works behind the scenes to streamline data exchange—is often more valuable than high-profile diagnostic tools.
Healthcare innovation must improve care delivery without creating new burdens. AI that achieves this earns clinician confidence. Mellin’s work situates AI adoption within a pragmatic framework of evaluation, safety, and real-world benefit, ensuring that the future of medicine is as efficient as it is advanced.
Follow Andrew on LinkedIn here.
42. Dr. Lynne Nowak — Building Educational Pathways for Clinicians
Dr. Lynne Nowak has focused on demystifying AI for clinicians by connecting technological advancement to practical, ethical use. As the Chief Data and Analytics Officer at Surescripts and a former Chief Medical Officer at Lark Health, Nowak is a seasoned internal medicine physician who bridges the gap between frontline care and complex data science. She consistently emphasizes that AI in healthcare must be paired with education, context, and clinical insight rather than adopted blindly (see: Lynne Nowak on Data and Patient Access).
Through her public commentary and leadership, Nowak highlights the importance of AI literacy—the ability for clinicians to understand not just what tools can do, but how concepts like pretraining and fine-tuning affect clinical outputs. She argues that physicians need structured, tiered competency frameworks to evolve from cautious users to responsible stewards of AI. This includes developing the "cognitive capability" to identify automation bias and evaluate AI against foundational medical knowledge.
In her work with organizations like the AMA Ed Hub and during industry forums like ViVE and HLTH, Nowak discusses how AI should be used to solve "high-friction" problems—such as complex record reviews and automated prior authorizations—while leaving empathetic, high-order reasoning to humans. Her perspective ensures that as technology scales, it remains anchored in the goal of reducing clinician burnout and improving health equity.
Healthcare technology can only improve care if clinicians are prepared to guide its use. Preparedness depends on thoughtful education. And thoughtful education depends on voices willing to bridge clinical expertise with emerging tools. Nowak’s work situates AI within that educational bridge—urging that clinician understanding be a prerequisite for responsible deployment.
Follow Lynne on LinkedIn here.
43. Dr. Christopher Longhurst — Academic Deployment and Validation
Dr. Christopher Longhurst has explored how digital transformation and AI must be integrated into clinical workflows to improve care delivery rather than add complexity. As the Chief Clinical and Innovation Officer at UC San Diego Health, Longhurst oversees clinical operations and digital strategy, positioning AI as a key tool for creating a "learning health system." He emphasizes that tools should support clinician decision-making while aligning with workflow and safety goals (see: Christopher Longhurst on Healthcare AI Integration).
In landmark 2025–2026 studies published in NEJM AI and JAMA Network Open, Longhurst has documented the "unreasonable effectiveness" of locally trained AI models. His work includes pioneering the use of large language models (LLMs) to automate hospital quality reporting and streamline the routing of patient messages in the EHR—demonstrating that while AI may not always save time, it significantly reduces cognitive burden and work exhaustion.
Longhurst is a vocal advocate for "Augmented Intelligence," arguing that AI should function as a second set of eyes rather than a replacement for human judgment. In his discussions with the National Academy of Medicine and on the No Priors podcast, he underscores that successful AI adoption depends on a solid digital infrastructure, rigorous "in vivo" validation, and maintaining the "human touch" in patient care.
Healthcare transformation depends not just on innovation but on integration. Integration depends on infrastructure and workflow alignment. And workflow alignment depends on leaders who understand both clinician needs and technology constraints. Longhurst’s work ensures that innovation supports clinicians where they work, keeping physicians "in the loop" to provide the empathy and ethics that AI cannot replicate.
Follow Christiopher on LinkedIn here.
44. Dr. Michael Blum — Operational Innovation at Scale
Healthcare AI depends on infrastructure that supports secure, real-world model validation, and Dr. Michael Blum has been building exactly that. As the Associate Vice Chancellor for Informatics and Chief Digital Transformation Officer at UCSF, Blum has spent decades integrating digital technologies—including AI and machine learning—into clinical care and health systems.
Blum is the CEO and Co-founder of BeeKeeperAI, a company spun out from UCSF’s Center for Digital Health Innovation (CDHI). He recognized that the primary bottleneck in healthcare AI is not the math, but the "data access problem"—developers cannot access high-quality patient data due to privacy risks, and health systems cannot share data without compromising sovereignty.
In his interview on the Navigating Forward podcast, Blum discussed how his "zero-trust" platform, EscrowAI, allows algorithm owners to bring their models to the data rather than moving the data to the model. This "sightless computing" ensures that neither the data steward nor the developer ever sees the other’s intellectual property, fulfilling a critical requirement for Coalition for Health AI (CHAI) standards.
Through recent collaborations with institutions like Mount Sinai and Morehouse School of Medicine, Blum’s work has demonstrated that privacy-preserving platforms can reduce the time to validate an AI model from years to weeks. His advocacy centers on the idea that traditional de-identification is no longer sufficient; instead, healthcare needs hardware-based security to enable the continuous, real-world testing required for safe clinical adoption.
Healthcare models are only as good as the data they train on, and real-world validation matters most. That validation depends on systems designed for both privacy and clinical rigor. Blum’s work situates clinical AI adoption within that practical framework of secure, responsible innovation.
45. Dr. Alistair Erskine — Analytics-Driven Decision Making
Data without insight changes nothing, and Dr. Alistair Erskine focuses on turning healthcare data into operational value. As the Chief Information Digital Officer at Highmark Health—a role he assumed in 2025 after leading digital strategy at Emory Healthcare—Erskine oversees the digital roadmap for one of the nation’s largest "payvidor" (payer-provider) organizations.
Erskine argues that for AI to be effective, it must be integrated into care delivery to support clinician decision-making and system efficiency, not just to meet technical benchmarks. In a 2026 update on standardizing AI-generated patient messages, he emphasized the need for structure and rigor to ensure that automated outreach and care navigation tools are safe and effective. His work often highlights the importance of data quality, governance, and transparency so that predictive models can be trusted by both clinicians and administrators (see: Putting AI into Practice).
Through initiatives like the Healthcare AI Challenge, Erskine advocates for putting "clinicians in the driver’s seat." He believes that analytics teams must partner deeply with stakeholders to build dashboards and early warning systems that align with workflow realities. His philosophy centers on the idea that insight is only useful when it informs action—such as ambient AI for prior authorizations—rather than simply reporting on the past.
Healthcare depends on data that leads to actionable understanding. Actionable understanding depends on interpretability and integration. And interpretability depends on leaders who connect analytics with clinical context. Erskine’s work positions AI and predictive models as pragmatic tools embedded in everyday decision support—where insight matters because it drives measurable improvements.
Follow Alistair on LinkedIn here.
46. Dr. Rasu Shrestha — Innovation Ecosystem Architect
Predictive tools mean little if they never reach the point of care, and Dr. Rasu Shrestha has spent his career building the bridges that bring analytics and AI into everyday clinical use. As the Enterprise Executive Vice President and Chief Innovation and Commercialization Officer at Advocate Health (the system formed by the merger of Atrium Health and Advocate Aurora Health), Shrestha is a primary architect of "intentional innovation."
Shrestha emphasizes that for AI to be successful, it must move from being a "shiny object" to a tool that humanizes care. In his 2024 and 2025 keynotes, including at the Institute for Experiential AI, he argues that the goal of AI should be to bring down the "artificial wall" of technology—like the EHR—that often stands between the clinician and the patient. He advocates for a shift from a "patient-centered" to a "person-centered" approach, leveraging AI to understand the stories and motivators behind the data.
Currently, Shrestha leads major collaborative efforts such as the CancerX Initiative, where he serves on the steering committee to apply digital innovation toward the national goal of reducing cancer death rates. His work focuses on "partnerships done right," where vendors, clinicians, and data scientists align their strengths to solve complex challenges like clinical trial acceleration and equitable care delivery.
Healthcare transformation requires tools that become part of real workflows. Real workflows depend on alignment between technology and human practice. And alignment depends on leaders who shape strategy around both. Shrestha’s work positions AI adoption inside a pragmatic ecosystem of partnership and continuous measurement, ensuring that technology serves as an enabler of health, hope, and healing.
47. Dr. Daniel Kraft — Exponential Medicine and Future Readiness
Dr. Daniel Kraft has spent years connecting the dots between exponential technologies and the future of medicine, including how artificial intelligence can augment clinical insight. As the founder and chair of NextMed Health (the evolution of Exponential Medicine), Kraft curates conversations at the intersection of AI, genomics, and digital health to help clinicians move from "sick care" to continuous, proactive "generative health."
Kraft emphasizes that AI should improve human decision-making rather than replace it. In his Medscape 2050 series and various TED talks, he points to a future where "digital exhaust" from wearables and AI-driven "co-bots" support earlier detection and personalized therapy. He argues that clinicians who use AI will ultimately replace those who do not, as these tools become essential for managing the sheer volume of modern medical data.
Kraft’s commentary, including his work for Forbes and his role as host of the Healthy Conversations podcast, stresses that AI’s impact must be measured by improved patient outcomes rather than just technical benchmarks. He advocates for a "cross-disciplinary" mindset, urging physicians to break out of clinical silos and collaborate with engineers and data scientists to build tools that respect safety, equity, and accountability.
Innovation in healthcare begins with informed clinicians and leaders. Informed leaders depend on credible frameworks for evaluation. And credible frameworks depend on voices willing to articulate how AI fits into clinical reality. Kraft’s work positions AI as a powerful catalyst for a shift toward "Stage 0" medicine—where disease is detected and intercepted before it starts.
Follow Daniel on LinkedIn here.
48. Dr. Tom Lawry — Strategic AI Frameworks for Healthcare
Dr. Tom Lawry has shaped how healthcare organizations think about strategic AI adoption by grounding emerging technologies in practical, ethical frameworks. As the Managing Director of Second Century Tech and the former National Director for AI for Health and Life Sciences at Microsoft, Lawry is a leading advisor to health and medical leaders worldwide, helping them navigate the transition to becoming "Intelligent Health Systems."
Lawry is the author of the best-selling books Hacking Healthcare: How AI and the Intelligence Revolution Will Reboot an Ailing System and the newly released Health Care Nation: The Future Is Calling and It's Better Than You Think. In his writing and keynote talks, he emphasizes that AI in healthcare should be evaluated not just for technical performance, but for its ability to humanize care and restore the "joy of medicine" by removing administrative friction for practitioners.
Through his work with global organizations, Lawry explores practical AI governance, highlighting the importance of responsible AI strategies and cross-functional collaboration. He argues that the biggest barrier to AI adoption is often not the technology itself, but the "leadership gap"—the need for a new set of competencies to manage the cultural shift toward a data-driven enterprise.
Innovation in healthcare must be tethered to strategy, safety, and clinician trust. Strategic adoption depends on governance and measurement. And governance depends on leaders willing to tie technology to care priorities, equity, and real outcomes. Lawry’s work positions AI adoption within that disciplined context—where promising tools are matched to meaningful clinical needs to reboot an ailing system.
49. Dr. Michael Howell — Quality Measurement and AI Integration
Healthcare tools must demonstrate clear benefit against established quality metrics, and Dr. Michael Howell has focused on exactly that intersection of analytics, outcomes, and responsible adoption. As the Chief Clinical Officer at Google, Howell leads a team of clinical experts focused on bringing the best of technology to the world of health, having previously served as the Chief Quality Officer at the University of Chicago Medicine.
In his foundational 2024 review for BMJ Quality & Safety, "Generative artificial intelligence, patient safety and healthcare quality," Howell provides a primer on how the shift from task-specific "AI 2.0" to foundation-model-driven "AI 3.0" can address recalcitrant problems in patient safety. He argues that while these models offer transformative capabilities—such as improving health literacy by tailoring complex medical notes for different audiences—they also introduce novel risks like hallucinations that must be managed through rigorous clinical governance.
Through his work on the "QUEST Perspective" (Quality, Equity, and Safety Together), Howell emphasizes that integrating analytics into quality frameworks is essential for reducing harm. In his JAMA podcast and 2026 "The Check Up" keynote, he articulates that AI should be viewed as a "powerful new teammate" in the healthcare team sport. He stresses that for AI to earn its place, it must be judged by measurable improvements in care delivery—such as easing clinician workloads or democratizing expertise—rather than theoretical performance alone.
Innovation must be tied to outcomes. Outcomes depend on measurement. And measurement depends on leaders who shape how data supports safer care delivery. Howell’s work positions AI and predictive tools within that disciplined framework of quality, safety, and measurable clinical impact, ensuring that the next frontier of medicine remains fundamentally patient-centered.
Follow Michael on LinkedIn here.
50. Dr. Andrew Ng — AI Democratization and Agentic Innovation
For artificial intelligence to achieve its full potential, it must be democratized through education and practical application, and Dr. Andrew Ng has spent his career at exactly that intersection of academic rigor and global accessibility. As the Founder of DeepLearning.AI and Landing AI, and a Managing General Partner at AI Fund, Ng leads the movement to empower individuals and enterprises to build an AI-powered future, having previously served as the founding lead of the Google Brain project and Chief Scientist at Baidu.
In his influential "AI Transformation Playbook" and more recent 2024 insights on Agentic Workflows, Ng provides a strategic primer on how the shift from supervised learning to agentic AI—where systems reason and iterate through tasks—can solve complex problems that traditional models cannot. He argues that while large language models are transformative, the next leap in productivity will come from "AI agents" that can plan and use tools autonomously, provided they are built on a foundation of responsible data practices and clear business logic.
Through his "Data-centric AI" movement, Ng emphasizes that the quality of data is more critical than the complexity of the code. In his frequent "The Batch" newsletters and global keynotes, such as his 2026 briefings on the "AI Revolution," he articulates that AI is "the new electricity," a general-purpose technology that will transform every industry. He stresses that for AI to be truly beneficial, it must be accessible to more than just big tech companies—advocating for "small data" solutions that allow local businesses and specialized fields to leverage automation.
Innovation must be accessible. Accessibility depends on education. And education depends on leaders who can demystify complex systems for the masses. Ng’s work positions machine learning within a framework of universal literacy and ethical implementation, ensuring that the next frontier of technology serves to augment human capability on a global scale.
Follow Andrew on LinkedIn here.
51. Dr. Fei-Fei Li — Human-Centered AI and Spatial Intelligence
For artificial intelligence to truly understand our world, it must move beyond processing pixels and text to comprehending the physical reality we inhabit, and Dr. Fei-Fei Li has spent her career at the vanguard of this evolution. As the Co-Director of the Stanford Institute for Human-Centered AI (HAI) and the Founder and CEO of World Labs, Li is pioneering the next frontier of "Spatial Intelligence"—the ability for AI to reason about and interact with the three-dimensional world. Often called the "Godmother of AI," she previously served as the Director of the Stanford AI Lab and Chief Scientist of AI/ML at Google Cloud.
In her seminal 2024 TED talk and subsequent 2025 landmark essay, "Spatial Intelligence: AI's Next Frontier," Li provides a primer on why the shift from language-centric models to "World Models" is essential for the future of technology. She argues that while Large Language Models (LLMs) have mastered human abstraction, the next leap in intelligence requires machines to understand physics, geometry, and occlusion. Her work at World Labs—which reached unicorn status within months of its 2024 launch—focuses on building AI that can generate and navigate physically consistent 3D environments, a capability she views as a "civilizational technology."
Through her leadership at Stanford HAI, Li emphasizes that the development of AI must be inextricably linked to human values. In her 2023 memoir, The Worlds I See, and her 2026 industry briefings, she articulates a vision where technology serves as an ethical partner in healthcare, education, and robotics. She argues that just as her creation of ImageNet provided the data foundation for the deep learning revolution, spatial intelligence will provide the architectural foundation for Embodied AI—allowing robots to move from controlled lab settings to assisting in complex, real-world human environments.
Innovation must be anchored in empathy. Progress depends on perception. And perception depends on leaders who view technology through a humanistic lens. Li’s work positions AI not as a replacement for human agency, but as a bridge between digital computation and physical reality, ensuring that as machines learn to "see" and "act," they do so in a way that fundamentally enhances the human experience.
Follow Fei-Fei on LinkedIn here.
52. Dr. Atul Butte — Translating Trillions of Data Points into Cures
For the vast oceans of medical data to truly serve humanity, they must be treated not as a storage burden but as a "digital gold mine" for discovery, and Dr. Atul Butte has spent his career as one of medicine’s premier "treasure hunters." As the inaugural Director of the Bakar Computational Health Sciences Institute at UCSF and the Chief Data Scientist for the entire University of California Health System, Butte oversees a massive data warehouse spanning six medical centers and millions of patients, transforming raw clinical records into life-saving insights.
In his groundbreaking work on "Data-Driven Systems Medicine," Butte provides a roadmap for how trillions of data points—from genomic sequences to electronic health records—can be used to "thaw" frozen discoveries. He argues that the shift from traditional hypothesis-driven research to unsupervised machine learning allows us to find novel patterns in disease phenotypes that human intuition might miss. His lab famously demonstrated this by using open-source data to identify that a common antidepressant could be repurposed to treat a specific type of lung cancer—launching clinical trials in a fraction of the usual time and cost.
Through his leadership of the UC system's Center for Data-driven Insights and Innovation (CDI2), Butte emphasizes that data is the ultimate "equalizer" in healthcare. In his recent 2025 and 2026 keynotes, including the celebrated "BCHSI@10" symposium, he articulated a vision of "AI as a Scalable Privilege," where high-end precision medicine is democratized through automation. He stresses that by aggregating data at scale, we can identify which treatments work for which specific populations in "real-world" time, moving past the limitations of traditional clinical trials to a more agile, evidence-based model of care.
Innovation must be rooted in evidence. Evidence depends on aggregation. And aggregation depends on leaders who view data as a public good. Butte’s work positions computational health at the very center of the clinical mission, ensuring that the wealth of information generated by today's patients becomes the foundation for the cures of tomorrow.
53. Dr. Kalie Dove-Maguire — Clinical Intelligence and Physician-Centric AI
For AI to truly revolutionize the bedside, it must transition from a digital burden to a clinical superpower, and Dr. Kalie Dove-Maguire is at the forefront of engineering that shift. As the President and Chief Product Officer at Evidently, and an Assistant Clinical Professor of Emergency Medicine at UCSF, Dove-Maguire bridges the gap between complex data science and the high-stakes reality of acute care. She leads the development of AI tools designed to solve the "data drowning" crisis—where clinicians spend more time navigating thousand-page charts than treating patients.
In her deep-dive appearance on the Offcall podcast episode, "Building Decision Support AI That Doctors Truly Love and Trust," Dove-Maguire provides a masterclass on why clinical judgment cannot be automated. Recounting a pivotal moment from her residency—where a critical detail of pulmonary hypertension was missed because it was buried deep in a transplant report—she argues that medical errors are often a failure of workflow, not of effort. Her work at Evidently focuses on "showing the work," building AI that doesn't just offer an answer, but cites its sources and surfaces conflicting data so physicians can exercise their lifelong skill of "seeking the truth."
Through her advocacy for clinician-led design, Dove-Maguire emphasizes that AI should be viewed as an "AI Resident"—a capable assistant that augments rather than replaces the human in the loop. In her 2026 strategic partnerships and keynotes, she articulates a vision where technology reduces the "cognitive tax" of administrative burdens while protecting the clinician's agency. She stresses that for AI to earn its place in the ER, it must be transparent and built with the primary goal of returning the doctor’s focus to the patient.
Innovation must be intuitive. Adoption depends on trust. And trust depends on leaders who have lived the challenges of the front lines. Dove-Maguire’s work positions clinical intelligence as the essential lubricant for a friction-heavy healthcare system, ensuring that as data grows exponentially, the human element of medicine remains central and supported.
Follow Kalie on LinkedIn here.
54. Dr. Jay Parkinson — Reimagining the Architecture of Modern Medicine
For healthcare to truly innovate, it must move beyond digitizing old workflows and begin designing entirely new ones, and Dr. Jay Parkinson has spent twenty years as medicine’s lead architect of the "new possible." As a pediatrician and serial entrepreneur, Parkinson has consistently predicted the industry's next move—from launching one of the first tech-enabled house call services in Brooklyn to founding Sherpaa, a pioneer in virtual primary care. Today, as the founder of Automate Clinic, he is bridging the gap between clinical nuance and artificial intelligence.
In his revealing interview on the Offcall podcast, "Innovation and AI-Driven Healthcare," Parkinson discusses how the iPhone—not the EHR—was the true inflection point for modern medicine. He argues that the majority of physical doctor visits add unnecessary friction and that the future of care lies in a "simple front door" similar to the user experiences of Uber or Google. At Automate Clinic, he is putting this philosophy into practice by training AI models to surface clinical consensus rather than just a single "right" answer, capturing the collective wisdom of physicians to handle the "weird" edge cases where standard guidelines often fail.
Through his work at the intersection of design and delivery, Parkinson emphasizes that "design is a form of protest." He articulates that in a fragmented system that often feels dehumanizing, a beautiful, seamless patient experience is a radical act of care. He stresses that for AI to be successful, it must be shaped by doctors who are "curious enough to break stuff," ensuring that automation serves to simplify access rather than add another layer of bureaucratic complexity.
Innovation must be human-centered. Human-centricity depends on design. And design depends on leaders who aren't afraid to reject the one-size-fits-all medical career path. Parkinson’s work positions clinical AI not as a replacement for the doctor, but as a structural redesign that restores freedom and creativity to the practice of medicine, ensuring that the next frontier of care is as intuitive as it is effective.
55. Yann LeCun — Building the Physics of Clinical Reasoning
For artificial intelligence to move beyond surface-level patterns and achieve true clinical reliability, it must transition from predicting words to understanding the world, and Yann LeCun is the architect of that shift. As a Turing Award winner and the Chief AI Scientist at Meta, LeCun is recognized globally as one of the "Godfathers of AI." While not a physician, his influence on medicine is foundational: his pioneering work on neural networks paved the way for modern medical imaging, and he is now focused on "World Models"—the reasoning engines he believes are the essential next step for safe, high-stakes healthcare.
In the featured On/Offcall briefing, "What Is a World Model in AI? AMI Labs' Yann LeCun and CEO Alex LeBrun Explain," LeCun provides a searing critique of current generative AI. He argues that while Large Language Models (LLMs) are impressive, they are essentially "stochastic parrots" that lack a basic understanding of causality and persistent reality. He posits that for AI to be truly useful in medicine, it must evolve into "World Models"—systems that reason and show judgment just like humans by simulating the physical and biological consequences of an action before it is taken.
Through his leadership at AMI Labs, a joint venture with Nabla CEO Alex Lebrun, LeCun is moving AI past the "ambient" era of simple documentation and into the era of augmented clinical judgment. He articulates that the future of medicine requires AI that doesn't just process text, but understands the underlying "world" of the patient—integrating continuous physiological signals and complex diagnostic logic. He stresses that these models are designed to support physician judgment, acting as a "reasoning teammate" that helps clinicians navigate the edge cases where generic guidelines often fail.
Innovation must be grounded in reasoning. Reasoning depends on a model of the world. And building that model depends on leaders who can see past the limitations of current chatbots toward a deterministic, reliable future. LeCun’s work positions machine learning as a core pillar of clinical safety, ensuring that as AI scales, it does so with an internal map of human biology that allows it to serve as a truly intelligent partner at the bedside.
56. Dr. Byron Crowe — Navigating Autonomous AI and the Future of Prescribing
For healthcare innovation to be meaningful, it must move beyond theoretical benchmarks toward a system of "graduated autonomy" that actually unburdens the clinician, and Dr. Byron Crowe is the leading strategist in that transition. As the Chief Medical Officer at Doctronic AI and an Assistant Professor of Medicine at Harvard Medical School, Crowe is at the center of the most provocative development in health tech: the first state-approved program in the U.S. allowing AI to autonomously renew prescriptions.
In his headline-making interview on the Offcall podcast, "Inside the First Autonomous AI Prescription Program in America," Crowe unpacks the reality of "AI-native care." He argues that the status quo of prescription refills—often processed with minimal physician review due to volume—is fundamentally broken and unsafe. At Doctronic, he helped pioneer a model that begins with human-in-the-loop oversight and graduates toward full autonomy only when the system has proven its reliability. He counters the fear of job loss with the concept of "doctor reassignment," where AI handles the routine documentation and refills, allowing physicians to focus on higher-level diagnosis and complex human connection.
Through his work as the lead author of the Society of General Internal Medicine’s (SGIM) position statement on generative AI, Crowe emphasizes that technical accuracy is only half the battle. In his 2026 briefings, he articulates the moral standard of "Careworthiness"—the idea that AI deployments should be judged by whether the clinicians who built them would stand behind every decision made. He stresses that liability is a question most companies sidestep, but one that Doctronic addresses head-on by integrating licensed physicians into every layer of the AI's "residency."
Innovation must be accountable. Accountability depends on clinical oversight. And oversight depends on leaders who are willing to bridge the gap between academic research and real-world implementation. Crowe’s work positions autonomous tools not as a replacement for the medical profession, but as a necessary redesign of the clinical workflow, ensuring that the "personal AI doctor" remains fundamentally anchored in human safety and physician-led governance.
Follow Byron on LinkedIn here.
57. Dr. Raj Narula — Architecting Virtual Neurology at Scale
For specialized medicine to truly serve all patients, it must transcend geographic limitations through high-fidelity digital infrastructure, and Dr. Raj Narula has spent his career building the bridges to make that possible. As the Founder and Chief Executive Officer of Sevaro, Narula leads one of the nation’s premier telestroke and teleneurology organizations, having established a model that delivers subspecialty expertise to rural and underserved hospitals with the same speed and precision as a physical academic medical center.
In his 2026 feature on the Offcall podcast, "Virtual Neurology at Scale: Raj Narula and Melanie Winningham on How Sevaro Is Transforming Rural Stroke Care”, Narula provides a blueprint for what he calls "Synchronous Specialty Care." He argues that in stroke care, where "time is brain," the bottleneck is rarely a lack of technology, but a lack of specialized eyes at the bedside. Through Sevaro, he has pioneered a platform that guarantees neurologist response times in under minutes, proving that virtual care is not a "second-tier" substitute but a fundamental redesign of the neurocritical care pathway. He posits that AI will soon act as the "triage engine" for these specialists, identifying the highest-risk patients before they even reach the CT scanner.
Through his leadership of Sevaro’s OneCall technology platform, Narula emphasizes that the "last mile" of healthcare is often the hardest to solve. In his recent keynotes, he articulates that the success of telehealth depends on "Frictionless Logistics"—ensuring that the data, the specialist, and the bedside nurse are aligned in a single, unified workflow. He stresses that for virtual neurology to be sustainable, it must be judged by measurable clinical outcomes—such as door-to-needle times for thrombolytics—rather than just the number of consultations logged.
Innovation must be equitable. Equity depends on distribution. And distribution depends on leaders who can scale human expertise through robust digital architecture. Narula’s work positions teleneurology as the frontline of a decentralized health system, ensuring that a patient’s zip code no longer determines their survival during a neurological emergency.
And Finally: Dr. Graham Walker — Building AI Where Physicians Actually Work
We would be remiss if we didn’t include Offcall’s co-founder, Dr. Graham Walker, to round out this list.
Much of the AI conversation happens in labs and boardrooms. Dr. Graham Walker’s work happens inside physician workflows. As co-founder of MDCalc, he helped digitize evidence-based clinical decision tools used by a majority of U.S. physicians. That foundation matters because AI does not enter medicine as a revolution. It enters through tools clinicians already trust. Adoption in healthcare is rarely explosive — it is incremental, layered onto existing systems.
Walker later co-authored The Physicians’ Charter for Responsible AI, a clinician-led framework outlining how AI should be deployed in healthcare. Rather than focusing on what AI can do, the Charter centers what it must not do: override physician oversight, operate without transparency, or introduce silent risk.
Through commentary shared on Offcall and on LinkedIn, Walker repeatedly emphasizes liability clarity, explainability, and workflow integration. AI must serve the cognitive process of clinicians — not distort it. Healthcare AI succeeds when it feels like an extension of clinical reasoning. That is the lens through which Walker approaches the future.
Follow Graham on LinkedIn and be sure to subscribe to Offcall’s weekly newsletter to follow more AI updates every week.
Offcall Team is the official Offcall account.
Comments
(0)
Join the conversation
See what your colleagues are saying and add your opinion.