AI in Medicine
Jump to navigation
Jump to search
SubPage-Summary of AI in Medicine
Audio File:AI versus experts in clinical diagnosis.m4a
Short video File:Future of Clinical AI.mp4
Clinical Diagnosis, Decision Support, and Prognosis
AI is boosting accuracy for clinicians, Philips North America CEO says
| Amina Niasse | Reuters | June 9, 2026
This article reports on a Philips-sponsored international survey suggesting that artificial intelligence is already improving clinician productivity, diagnostic accuracy, and workflow efficiency. Doctors surveyed said AI helped identify possible medical errors, reduce costs, and allow clinicians to see more patients, while also warning that training remains uneven and that most serious clinical decisions are still being handled by humans.
How good are ‘AI doctors’ — and will they take over medicine?
| Mariana Lenharo | Nature | June 3, 2026
This Nature article examines the growing debate over whether AI systems can function as diagnostic assistants or even partial substitutes for physicians. It discusses clinical trials, expert skepticism, and the possibility that future medical care may involve a three-way relationship among patient, doctor, and AI rather than a simple replacement of human clinicians.
As artificial intelligence shows off diagnostic chops, scientists reckon with the way forward
| Katie Palmer | STAT | April 30, 2026
This article covers a prominent study in Science showing that an OpenAI large language model performed strongly on diagnostic and clinical reasoning tasks, including experiments based on real emergency department cases. The article balances the impressive results with calls from physicians and researchers for rigorous clinical trials before such systems are trusted in real-world care.
A systematic review and meta-analysis of diagnostic performance comparison between generative AI and physicians
This peer-reviewed review analyzed 83 studies comparing generative AI systems with physicians on diagnostic tasks. It found that AI showed promise and sometimes performed similarly to non-expert physicians, but overall remained inferior to expert physicians, highlighting both the potential and the limitations of using generative AI for medical diagnosis.
| Marty Stempniak | Radiology Business | March 10, 2026
This article reports that ECRI identified the AI diagnostic dilemma as the top patient safety concern for 2026. It focuses on the risk that AI diagnostic systems may improve speed and accuracy in some cases while also introducing new dangers through bias, poor validation, overreliance, and errors that clinicians may not easily detect.
Advanced AI holds promise for high-stakes healthcare, studies show
| Andrea Fox | Healthcare IT News | May 6, 2026
This article discusses two high-stakes uses of medical AI: diagnostic reasoning by advanced language models and early cancer detection through imaging analysis. It is useful because it shows both the promise of AI in emergency medicine and the need for careful clinical validation, infrastructure, and monitoring before these systems are broadly deployed.
How to create “humble” AI
| Anne Trafton | MIT News | March 24, 2026
This article describes an MIT-led effort to design medical AI systems that express uncertainty rather than giving overconfident recommendations. It is useful because medical AI mistakes can become more dangerous when clinicians or patients treat AI as authoritative, making humility, uncertainty, and collaboration important design goals.
AI: Algorithm that performs prescription renewals “better than human clinicians” is allowed to make decisions in Sweden
This article reports on an AI system in Sweden that was legally allowed to participate in clinical decisions about prescription renewals. It is useful because prescription renewal sits between administration and clinical judgment, making it a strong example of how AI is beginning to cross into legally significant medical decision-making.
The algorithm will see you now? Patients say not without a doctor nearby
| Jared Wadley | Medical Xpress | March 5, 2026
This article reports on a University of Michigan and Michigan State University study showing that patients are more willing to accept medical AI when it performs well, uses representative data, has regulatory or institutional approval, and includes a clinician in the loop. It is useful because it shows that patient trust depends not only on technical accuracy, but also on human oversight and institutional safeguards.
Can AI help decide when to see a doctor? Study says not yet
| Andrew Zinin | Medical Xpress | February 9, 2026
This article covers research showing that people using large language models to interpret symptoms and decide whether to seek care did not perform better than people using ordinary internet search. It is useful as a cautionary source because it shows that strong benchmark performance by AI systems does not necessarily translate into safe or effective real-world use by patients.
Reliability of LLMs as medical assistants for the general public: a randomized preregistered study
| Adam M. Bean and colleagues | Nature Medicine | 2026
This peer-reviewed study tested whether large language models could help members of the public identify medical conditions and choose an appropriate course of action. It found that while the models performed well when tested directly, ordinary users did not get reliable help from them, highlighting the importance of real-world human testing before deploying AI medical assistants to the public.
Can AI help predict which heart-failure patients will worsen within a year?
| Alex Ouyang | MIT News | March 12, 2026
This article reports on a deep-learning model developed by MIT, Mass General Brigham, and Harvard Medical School that forecasts whether a heart-failure patient’s condition may worsen within a year. It is useful because it shows AI being used for prognosis and follow-up planning, not simply diagnosis.
Documentation, Medical Records, Scribes, and Hospital Workflow
After hospitals, patients get a turn to bring AI into the doctor’s office
| Katie Palmer | STAT | June 4, 2026
This article looks at the spread of AI-powered medical scribes from hospitals and doctors’ offices into patient-controlled tools. It explores how patients are beginning to use apps that record, summarize, and interpret medical visits, raising questions about accuracy, privacy, physician trust, and whether AI can help people better understand their own care.
The medical AI revolution requires rethinking health care’s architecture
| Freddy Abnousi and Celina Yong | STAT | April 16, 2026
This article argues that medical AI will be limited if it only has access to chart notes, lab values, images, and short clinic encounters. The authors suggest that truly transformative AI will require a broader health care architecture that captures lived experience, daily function, symptoms over time, and context that is often missing from formal medical records.
AI agents are rapidly spreading in health care, but validation is lacking
| Casey Ross | STAT | March 11, 2026
This article examines the rapid spread of AI agents from major health technology companies, including systems designed to automate clinical and administrative tasks. It emphasizes concerns that many tools are being marketed faster than they are being validated, with patients often left out of development, testing, and safety evaluation.
Nvidia Is Developing an AI Healthcare Model With Startup Abridge
| Isabelle Bousquette | The Wall Street Journal | June 11, 2026
This article covers Nvidia’s partnership with Abridge to build a healthcare-specific AI model for clinical conversations. The model is intended to improve medical documentation and may eventually support real-time clinical decision-making, showing how major technology companies are moving deeper into everyday medical workflows.
AI in medical practice: doctors’ perspective on the benefits, challenges and facilitators of artificial intelligence scribe use
This study examines doctors’ views of AI scribes, including perceived benefits such as faster documentation and improved patient interaction, as well as concerns about errors, privacy, medico-legal responsibility, overreliance, and loss of clinical control. It is useful for understanding how physicians themselves see the practical rollout of AI in medicine.
More than 80% of physicians use AI professionally: AMA survey
| American Medical Association | AMA | March 12, 2026
This article summarizes an AMA survey finding that physician use of AI has grown sharply since 2023, with common uses including medical research summaries, discharge instructions, care plans, documentation, chart summaries, patient message drafts, translation, and assistive diagnosis. It is useful because it shows that AI is already becoming part of routine medical work, especially in documentation and information management.
Doctors increasingly see AI scribes in a positive light: But hiccups persist
| Medical Xpress | Medical Xpress | February 4, 2026
This article examines the rapid spread of ambient AI scribes that listen to clinical visits and draft notes for physicians. It is useful because it shows why many doctors see AI documentation tools as a relief from electronic health record burden, while also raising concerns about vigilance, errors, equity, and whether smaller practices can afford the technology.
Why some hospitals are making their own ChatGPTs for patient records
| Brittany Trang | STAT | January 28, 2026
This article reports that Stanford Health Care and Penn Medicine have developed internal AI chat tools that let clinicians query and summarize patient records. It is useful because it shows a practical hospital use case for generative AI: helping doctors navigate long, complicated electronic health records before and during patient care.
Large AI scribe study finds modest time savings, inconsistent benefits
| Brittany Trang | STAT | April 1, 2026
This article covers a large study of AI medical scribes involving about 1,800 clinicians across five academic medical centers. It found that the tools saved some documentation time, but the benefits were uneven, making it useful for understanding why AI scribes may help some doctors more than others.
No, I don’t want an AI scribe to write my pulmonologist’s note
| Aliaa Barakat | STAT | April 15, 2025
This opinion article argues that physician-written notes are not just administrative paperwork, but part of clinical reasoning and the doctor-patient relationship. It is useful as a skeptical counterpoint to AI scribe enthusiasm, emphasizing what may be lost when documentation is outsourced to automated systems.
AI could ease the burden of hospital discharge summaries
| Christina Hernandez Sherwood | Stanford Medicine | June 8, 2026
This article reports on a small pilot study of a Stanford Medicine-developed AI tool that helps doctors create hospital discharge summaries. It is useful because discharge summaries are time-consuming but clinically important documents, and the study found the tool appeared safe while also reducing physician burnout.
| John Halamka and Paul Cerrato | Mayo Clinic Platform | May 29, 2026
This article looks at how AI-enhanced systems could help patients and clinicians manage the complicated logistics of modern health care. It is useful because it treats medical AI not only as a diagnostic tool, but also as a guide through scheduling, records, referrals, insurance, follow-up care, and other everyday barriers that make the health system hard to navigate.
4 Health Systems Transforming Care with AI
| American Hospital Association | AHA Center for Health Innovation | May 12, 2026
This article gives examples of health systems using AI for radiology, virtual nursing, fall prevention, clinical workflow, and patient care operations. It is useful because it shows how AI is already being embedded into hospital systems, not just tested in research labs or discussed as a future possibility.
Patient-Facing AI, Trust, Mental Health, and Misinformation
Stanford Health Care brings patients into decisions on AI tools
| Brittany Trang | STAT | May 27, 2026
This article describes how Stanford Health Care has been asking patient panels to weigh in before deploying new AI tools. The piece is useful because it focuses less on technical performance and more on public trust, consent, patient priorities, and the need to include patients in decisions about how medical AI is tested and introduced.
Generative artificial intelligence-driven chatbots and health misinformation
| N. B. Tiller and colleagues | BMJ Open | April 2026
This article examines the risks of generative AI chatbots spreading inaccurate or misleading health information. It is useful because many patients now use chatbots before or after seeing clinicians, creating new challenges for medical advice, public health communication, and the relationship between patients and professionals.
Advertising to the distressed: The commodification of mental health data in AI chatbots
| Nicole Gross and Hannah van Kolfschooten | Journal of Medical Ethics Blog | April 14, 2026
This article raises ethical concerns about people using AI chatbots for emotional support and mental health conversations. It argues that deeply personal disclosures may become commercial data, creating risks around privacy, manipulation, behavioral advertising, and the exploitation of vulnerable users seeking help.
From Advice to Action — Real-World Behavior of Patients Using AI Health Advice
| F. Cotte and colleagues | NEJM AI | April 2026
This article studies how people act after receiving AI-generated health advice. It is useful because the real-world effects of medical AI depend not only on whether an answer is accurate, but also on whether patients understand it, trust it, act on it appropriately, or use it in ways that change their relationship with clinicians.
Mental Health, Optimism, and the AI Universe
| Mayo Clinic Platform | Mayo Clinic Platform | January 13, 2026
This article discusses the promise and limits of generative AI in mental health support. It is useful because mental health is one of the areas where patients may be most tempted to use AI directly, while also being one of the areas where trust, safety, emotional vulnerability, and professional oversight are especially important.
Meeting an AI doctor before a real-life consultation can improve cancer patients' understanding and reduce stress
| European Society for Radiotherapy and Oncology | Medical Xpress | May 16, 2026
This article reports on research presented at ESTRO 2026 suggesting that cancer patients who interacted with an AI avatar doctor before meeting a real consultant felt more informed and less anxious. It is useful because it presents AI as a patient-preparation tool, helping people understand radiation treatment and arrive at appointments ready to ask better questions.
Public trust in AI in health care is slipping, survey finds
| Ohio State University Medical Center | Medical Xpress | April 7, 2026
This article reports that public openness to AI in health care has declined in a national survey, even as AI tools become more common in medical settings. It is useful because it shows that the future of AI in medicine depends not only on technical success, but also on public confidence, transparency, and whether patients believe AI is being used appropriately.
Patients have reservations about physicians who use artificial intelligence, according to study
| Julius-Maximilians-Universität Würzburg | Medical Xpress | July 21, 2025
This article reports on a study finding that patients rated physicians as less competent, trustworthy, and empathetic when told the doctor used AI, even for administrative purposes. It is useful because it highlights a social barrier to medical AI adoption: patients may worry that doctors will blindly trust algorithms or become less personally engaged in care.
Can Patients Trust Chatbots to Manage and Interpret Their Medical Data?
| John Halamka and Paul Cerrato | Mayo Clinic Platform | April 20, 2026
This article examines the risks and potential benefits of patients using AI chatbots to collect, organize, and interpret their own medical records. It is useful because it focuses on a fast-growing patient-facing use of AI, where convenience must be weighed against privacy, accuracy, security, and the danger of misunderstanding complex clinical information.
Impact of AI misinformation on diagnostic accuracy and confidence calibration in novice medical students
This randomized trial studied how AI-generated misinformation affects novice medical students. It found that misleading AI explanations significantly harmed diagnostic accuracy, while correct explanations did not provide an equivalent benefit, making it an important warning about how persuasive but wrong AI output can distort medical learning and clinical reasoning.
Regulation, Governance, Liability, Transparency, and Safety Oversight
AI doctors should be licensed. Here’s a framework to do that
| Alon Bergman | STAT | May 11, 2026
This opinion article argues that autonomous clinical AI systems should be licensed and monitored in ways similar to human clinicians. Using Utah’s Doctronic pilot as an example, it calls for competency testing, supervised deployment, post-market monitoring, and federal standards to avoid a confusing patchwork of state rules.
Artificial Intelligence-Enabled Medical Devices
| U.S. Food and Drug Administration | FDA | 2026
This FDA resource lists AI-enabled medical devices authorized for marketing in the United States and explains the agency’s role in reviewing safety and effectiveness. It is a key source for understanding how AI is already entering medicine through regulated tools, especially in imaging, diagnostics, monitoring, and clinical decision support.
Show us the evidence for the value of medical AI
| Ravi B. Parikh and Isaac S. Kohane | Nature Medicine | April 21, 2026
This Nature Medicine commentary argues that health systems need stronger evidence that AI tools actually improve patient outcomes, clinician work, or system costs. It is a valuable counterbalance to optimistic AI coverage because it asks whether medical AI produces measurable value after deployment, not just impressive results in controlled tests.
Why conversations around health AI may be evolving beyond hype
| Brittany Trang | STAT | April 29, 2026
This article discusses a shift in the health AI conversation from broad excitement toward more serious questions about evidence, implementation, payment, safety, and patient benefit. It is useful for framing AI in medicine as a maturing field where the key issue is no longer whether tools are impressive, but whether they work responsibly in real health systems.
Clinical AI Has Boomed. A New Stanford-Harvard State of Clinical AI Report Shows What Holds Up in Practice.
| Stanford Medicine | Stanford Medicine | January 15, 2026
This article summarizes findings from the Stanford-Harvard State of Clinical AI report, emphasizing that more than 1,200 AI-enabled medical tools have already been cleared by the FDA while many claims still need real-world validation. It is useful for understanding the scale of clinical AI adoption and the gap between promising announcements and proven clinical impact.
A Collaborative Best Practice Guide for Promoting AI Vendor Transparency
| S. Kpodzro and colleagues | NEJM AI | May 2026
This article presents guidance for improving transparency from health AI vendors. It is useful because hospitals and health systems often buy AI tools from outside companies, and meaningful oversight requires clear information about training data, intended use, performance, limitations, bias, monitoring, and responsibility after deployment.
Doctors and NHS could be sued for mistakes made by AI tools, report warns
| Denis Campbell | The Guardian | June 9, 2026
This article reports on warnings from the Medical Protection Society that doctors and the NHS could be held legally responsible when AI tools make harmful diagnostic or treatment mistakes. It is useful because it focuses on liability, accountability, and the legal gap between fast-moving medical AI systems and slower-moving malpractice and product safety law.
With AI increasingly part of care, transparency and quality are musts
| American Medical Association | AMA | June 2026
This article describes AMA policy efforts to keep evidence-based care, transparency, and quality standards at the center of medical AI adoption. It is useful because it frames AI as an “augmented intelligence” tool that should enhance physician judgment rather than replace it, while warning that decision-support systems need strong oversight.
Request For Public Comment: Measuring and Evaluating Artificial Intelligence-enabled Medical Device Performance in the Real-World
| U.S. Food and Drug Administration | FDA | September 30, 2025
This FDA request for public comment focuses on how to measure and monitor AI-enabled medical devices after they are deployed in real clinical settings. It is useful because it addresses performance drift, data changes, bias, safety monitoring, and the need for ongoing real-world evaluation rather than relying only on premarket testing or static benchmarks.
Regulation of clinical Artificial Intelligence in the Age of LLMs
| C. Tan | npj Digital Medicine | 2026
This article examines how large language models fit into existing rules for clinical decision-support systems and medical devices. It is useful because it separates narrower AI tools designed for specific clinical indications from more general-purpose AI systems that may need new regulatory approaches.
Advancing healthcare AI governance through a resource-aware framework
| R. Hussein and colleagues | npj Digital Medicine | 2026
This article reviews health care AI governance frameworks and argues that many assume resources that smaller hospitals and clinics may not have. It is useful because it focuses on practical governance, including how real health systems can monitor, evaluate, and safely deploy AI without creating standards only large institutions can meet.
More Fragmented, More Complex: State Regulation of AI in Health Care
This article examines how U.S. states are creating separate rules for health care AI, producing a more complex and fragmented regulatory landscape. It is useful because hospitals, doctors, software companies, and patients may face different standards depending on where an AI tool is used.
AI Agents, Automaticity, and Value Alignment in Health Care
| R. A. Taylor | NEJM AI | 2025
This article discusses the rise of AI agents that can perform increasingly autonomous tasks in health care. It is useful because it focuses on value alignment: whether AI systems act in ways consistent with patient welfare, clinician judgment, medical ethics, and institutional goals.
Artificial Intelligence in Software as a Medical Device
| U.S. Food and Drug Administration | FDA | March 25, 2025
This FDA page explains how artificial intelligence and machine learning are being used in software-based medical devices. It is useful as a regulatory background source because it shows how the FDA thinks about AI tools that learn from health care data, support clinicians, and may change over time after deployment.
Artificial Intelligence-Enabled Device Software Functions: Lifecycle Management and Marketing Submission Recommendations
| U.S. Food and Drug Administration | FDA | January 7, 2025
This FDA draft guidance gives recommendations for companies submitting AI-enabled medical device software for review. It is useful because it addresses lifecycle management, risk assessment, safety, effectiveness, bias, transparency, and the evidence manufacturers should provide before AI medical tools are marketed.
States Continue Efforts to Regulate AI in Healthcare: A Review of Legislation Passed in 2026
| Dan M. Silverboard, Jacob Anthony Barrera, and Isabella Ryan | Holland & Knight | May 26, 2026
This article reviews state laws passed in 2026 that regulate AI in health care, including rules on insurer decision-making, prior authorization, autonomous clinical decisions, patient disclosure, informed consent, behavioral health chatbots, and AI companions. It is useful because it shows that health AI regulation is developing state by state while Congress has not yet created one national framework.
Ethical AI in Radiology: Performance, People, and Post-market Responsibility
| Alex Perkins | European Medical Journal | April 17, 2026
This article summarizes a European Congress of Radiology session arguing that ethical AI in radiology cannot be judged by accuracy alone. It is useful because it emphasizes post-market monitoring, cybersecurity, regulation, workflow design, and human factors as essential parts of safe medical AI deployment.
MIT scientists investigate memorization risk in the age of clinical AI
| Alex Ouyang | MIT News | January 5, 2026
This article explains research on how to test whether clinical AI models memorize and expose sensitive patient information from training data. It is useful because privacy is one of the central risks of medical AI, especially when models are trained on de-identified health records that may still contain patterns or details that could be leaked.
Drug Discovery, Biomedical Research, Clinical Trials, and New Medical Products
Takeda's AI-crafted psoriasis pill tops Bristol Myers' Sotyktu in head-to-head trial
| Reuters | Reuters | June 11, 2026
This article reports that Takeda’s experimental once-daily psoriasis pill, developed with the help of artificial intelligence, outperformed Bristol Myers Squibb’s Sotyktu in a late-stage trial. The story is useful because it gives a concrete example of AI-assisted drug development moving beyond theory and into a real clinical product pipeline.
After seeing Parkinson's take his father, this Googler found a new mission: teaching AI to chase cures
| Hugh Langley | Business Insider | June 11, 2026
This article profiles Google DeepMind researcher Vivek Natarajan and his work on medical AI tools such as Med-PaLM, AMIE, Co-Clinician, and Co-Scientist. It is especially useful for understanding how AI is being developed not just for diagnosis, but also for scientific hypothesis generation, drug discovery, and accelerating biomedical research.
Are 'AI co-scientist' tools actually useful for scientists?
| Brittany Trang | STAT | May 21, 2026
This article examines the promise and limitations of AI “co-scientist” systems designed to help researchers generate hypotheses and explore biomedical questions. It is useful because it pushes past hype and asks whether these tools are genuinely helping scientists or mainly producing outputs that still require heavy human judgment.
Artificial intelligence accelerates drug discovery and enhances commercialization efficiency
This open-access review examines how AI is being used across drug discovery, preclinical development, clinical studies, regulatory filing, pharmacovigilance, marketing, and commercialization. It is useful as a broad scholarly source on how AI may shorten or improve multiple stages of pharmaceutical development.
How artificial intelligence could reshape the medical technology industry
| M. Shafieian | BMJ Innovations | May 26, 2026
This article examines how AI could change the medical technology industry by reducing research and development costs, accelerating innovation, and changing the competitive landscape. It is useful because it looks beyond individual tools and asks how AI may alter the economics of medical devices, diagnostics, and health technology development.
Artificial Intelligence for Drug Development
| U.S. Food and Drug Administration | FDA | 2026
This FDA resource explains how artificial intelligence is being used throughout the drug development life cycle, including nonclinical research, clinical trials, postmarketing surveillance, manufacturing, digital health technologies, and real-world data analysis. It is useful as a regulatory source showing that AI is not only changing diagnosis, but also the way medicines are discovered, tested, monitored, and manufactured.
Accelerating AI innovation in healthcare: real-world clinical data and Mayo Clinic Platform
| Y. Yu and colleagues | npj Digital Medicine | 2026
This article describes how the Mayo Clinic Platform supports AI research through multi-institutional, de-identified clinical data and analytic tools. It is useful because it shows how high-quality data infrastructure is becoming central to developing, testing, and validating medical AI systems.
Considerations for the Use of Artificial Intelligence to Support Regulatory Decision-Making for Drug and Biological Products
| U.S. Food and Drug Administration | FDA | January 6, 2025
This FDA guidance explains how AI-generated information may be used to support regulatory decisions about drugs and biological products. It is useful because it shows how AI is moving into clinical trials, drug safety, quality evaluation, and evidence generation for new medical products.
Meeting the Challenge of Clinical Research in a Digital World
| John Halamka and Paul Cerrato | Mayo Clinic Platform | March 3, 2026
This article discusses how electronic health records and AI can be used to emulate clinical trials, predict disease progression, and study treatment effects. It is useful because it shows how AI may change medical research itself, making it possible to generate evidence from real-world clinical data when traditional trials are slow, expensive, or impractical.
Using synthetic biology and AI to address global antimicrobial resistance threat
| Alex Ouyang | MIT News | February 11, 2026
This article reports on an MIT research project using synthetic biology and generative AI to develop programmable antibacterials against antimicrobial resistance. It is useful because it shows AI being applied to one of the most serious global health threats: the rise of drug-resistant infections.
AI-generated sensors open new paths for early cancer detection
| Anne Trafton | MIT News | January 6, 2026
This article describes an AI model that designs peptide-coated nanoparticles capable of detecting cancer-associated protease activity and producing signals that may be read through a urine test. It is useful because it shows AI contributing to early cancer detection through new biological sensor design, rather than only through software analysis of existing scans or records.
Specialty Medicine: Oncology, Radiology, Surgery, Rehabilitation, and Primary Care
Artificial Intelligence Applications in Medical Devices for Rehabilitation
| Authors not listed in search result | Journal of Medical Internet Research | March 4, 2026
This peer-reviewed article reviews how AI is being used in medical devices for rehabilitation, including patient monitoring, assistive technologies, movement analysis, and personalized therapy. It broadens the topic beyond diagnosis and drug discovery by showing how AI may affect recovery, disability care, and daily functioning.
Artificial intelligence in primary care: frameworks, implementation, and safety
| L. N. Allen and colleagues | The Lancet Primary Care | 2026
This article reviews how artificial intelligence is being introduced into primary care, including documentation, triage, diagnosis support, risk prediction, and workflow management. It is useful because primary care is one of the places where AI may have the broadest impact, but also where safety, equity, patient trust, and clinician workload must be handled carefully.
As AI enters the operating room, reports arise of botched surgeries and misidentified body parts
| Chad Terhune and Robin Respaut | Reuters | February 9, 2026
This investigation examines safety reports, lawsuits, and regulatory concerns involving AI-enabled medical devices, including surgical navigation systems and diagnostic tools. It is useful as a cautionary source showing that AI in medicine is not only about promising software, but also about patient injuries, device recalls, FDA oversight, and accountability when systems fail.
AI Avalanche Is Forcing Healthcare to Reimagine Future of Surgery
| Lenworth M. Jacobs Jr. | American College of Surgeons Bulletin | January 2026
This article discusses how AI, robotics, autonomous systems, and remote surgery are pushing surgeons and hospitals to rethink clinical responsibility, safety standards, and governance. It is useful because it frames surgical AI as a systems problem, where infrastructure, oversight, training, and professional judgment matter as much as the technology itself.
Why oncology is becoming healthcare AI's toughest test
| Bill Siwicki | Healthcare IT News | June 2026
This article argues that oncology is one of the most difficult areas for trustworthy AI because cancer care depends on complex patient histories, rapidly changing treatment standards, and fragmented electronic health records. It is useful because it shows why good AI performance in demonstrations may not be enough when real clinical data are messy, incomplete, and highly consequential.
The role of AI in oncology: present applications and future directions
| A. Weitzner and colleagues | npj Precision Oncology | 2026
This article reviews how AI is being used in cancer care, including imaging, pathology, clinical trial matching, risk prediction, and treatment planning. It is useful because oncology is one of the highest-stakes areas for medical AI, where small errors can affect diagnosis, therapy selection, and patient survival.
Many Brain Tumors May Yield to AI Algorithms
| John Halamka and Paul Cerrato | Mayo Clinic Platform | June 4, 2026
This article discusses how AI algorithms may help classify and guide treatment decisions for meningiomas, which account for a large share of central nervous system tumors. It is useful because it shows how AI may support precision diagnosis in brain tumors, especially when classification systems become more detailed and difficult for clinicians to apply consistently.
Can AI Models Improve Diagnosis of Pancreatic Cancer?
| John Halamka and Paul Cerrato | Mayo Clinic Platform | March 12, 2026
This article discusses how AI models may help improve earlier detection of pancreatic cancer, one of the most difficult cancers to diagnose early. It is useful because it shows how AI could be especially valuable where human detection is limited by subtle symptoms, rare disease patterns, and the need to find warning signs before cancer becomes advanced.
Deepfake X-Rays Fool Radiologists and AI
| RSNA | Radiological Society of North America | March 24, 2026
This article reports on research showing that AI-generated “deepfake” X-ray images can fool both radiologists and multimodal AI systems. It is useful because it adds a cybersecurity and medical-record integrity angle to the AI-in-medicine debate, showing that generative AI may create new risks for imaging, diagnosis, fraud, and clinical trust.
Big-Picture Trends, Costs, Workforce, Equity, and Global Health
Cheap AI chatbots transform medical diagnoses in places with limited care
| Chris Simms | Nature | February 6, 2026
This Nature article reports on studies in Rwanda and Pakistan showing that AI chatbots may help support diagnosis in under-resourced clinics. The article is especially useful for understanding AI in medicine as a global health tool, where the issue is not replacing doctors but helping overstretched health systems provide more consistent care.
Who will pay for AI in health care? 3 trends to watch in 2026
| Katie Palmer | STAT | January 2, 2026
This article focuses on the economics of medical AI, including reimbursement, value-based care, and whether patients may be asked to pay directly for some AI-enabled services. It is useful for understanding why adoption of AI in medicine depends not only on accuracy or innovation, but also on payment systems, incentives, and proof of clinical value.
How AI is making health care even less affordable
| Caitlin Owens | Axios | June 12, 2026
This article argues that AI may initially make health care more expensive rather than cheaper, especially where AI tools help providers document more services and increase billing complexity. It is useful because it connects medical AI to the incentives of the health care payment system, showing that efficiency does not automatically translate into lower costs for patients.
Global advances in health artificial intelligence
| The Lancet | The Lancet | June 2026
This article reviews global progress in health artificial intelligence, including clinical applications, patient engagement, scientific discovery, and ethical concerns. It is useful as a broad overview of how AI is moving into health systems around the world while emphasizing that AI should support clinical judgment rather than replace patient-centered medical care.
Reconciling how clinical reasoning is learned in the age of artificial intelligence
| Andrew Y. Ong | npj Digital Medicine | 2026
This article discusses how AI may change medical education and the way clinical reasoning is taught. It is useful because it examines both the promise of supervised AI use for learning and the danger that plausible AI misinformation may distort how students develop diagnostic judgment.
Are AI scribes actually driving higher health care costs?
| Brittany Trang | STAT | April 8, 2026
This article looks at whether AI scribes may increase health care spending by producing more detailed documentation that supports higher billing. It is useful because it questions the assumption that AI documentation tools automatically save money, showing how medical AI can interact with the financial incentives of the health care system.
AI agent in healthcare: applications, evaluations, and future challenges
| L. Zhao and colleagues | npj Health Systems | 2026
This review examines the rise of AI agents in health care, including applications in diagnosis support, clinical decision-making, medical report generation, patient-facing chatbots, hospital operations, and medical education. It is useful because it explains the next step beyond simple chatbots: AI systems that can plan, use tools, and act across complex medical workflows.
A quantitative analysis of global AI medical studies: gaps in geography, disease focus, and evaluation
| Q. Wang and colleagues | npj Digital Medicine | 2026
This article analyzes global medical AI research and identifies uneven coverage across countries, diseases, and clinical settings. It is useful because it shows that the medical AI evidence base may be geographically and medically skewed, raising questions about whether models tested in one setting will work safely elsewhere.
The Utility of Generative AI in Advancing Global Health
| H. Akbarialiabad and colleagues | NEJM AI | 2025
This article explores how generative AI could improve health care access and quality in low- and middle-income countries. It is useful because it frames medical AI as a global health tool, while also warning that deployment must account for language, infrastructure, bias, safety, and local clinical realities.
The big ideas from Stanford Health AI week
| Stanford Medicine | Stanford Medicine | June 2026
This article summarizes major themes from Stanford Health AI Week, including patient collaboration, trust-building, responsible deployment, AI in mental health, medical imaging, life sciences, and translational medicine. It is useful as a broad overview of how academic medicine is trying to move AI from impressive demonstrations into safe, trusted clinical practice.
Is AI Threatening Physicians' Livelihood?
| John Halamka and Paul Cerrato | Mayo Clinic Platform | March 3, 2026
This article explores whether medical AI is a threat to doctors’ jobs or a tool that can support better care. It is useful because it addresses a central concern among clinicians: whether AI will replace professional judgment or instead become a partner that reduces administrative burden and improves decision support.
