Friday Roundup: vibe coding in research; frameworks for healthcare AI use; tracking diffusion of bogus citations; rethinking regulatory approaches for AI; existential doubt sown by AI; immunotherapy trial for glioblastoma; personal info surfaced by AI search; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: cooperating AI agents work effectively but alignment suffers; alarm as parents refuse vitamin K shots; automated paper review; “AI paradox” affects uptake; humans still beat AIs on creativity; journals face AI slop at both ends of peer review; more: aihealth.duke.edu/ai-health-ro...

The Patient-Centered Outcomes Research Institute (PCORI) has announced a funding award for Duke Health's Hypertension Management Acceleration Program (HTN MAP), which aims to improve blood pressure control in communities served by Duke Health: aihealth.duke.edu/tackling-hyp...

Friday Roundup: cursing AI agent runs amok, deletes database; Anthropic researchers: incoherence a besetting challenge in larger models; FDA announces real-time trial data monitoring; automation bias in docs using LLMs; licensure model for clinical AI systems; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: favoring owls and other signs of AI misalignment; sifting through health queries to Copilot; gamified ED triage training; LLMs, EHR notes & accountability; progress with JEPA world models; AI ushers in new era of risks for health system data, more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: AI Index shows public, expert views of AI diverging; telling scary stories about AI; rethinking cancer care; dubious data may have trained predictive models; ancient origins for antiviral tools; are preprints repeating open-access problems?; more: aihealth.duke.edu/ai-health-ro...

Thursday, 4/16, 11:30 AM: Duke Center for Computational & Digital Health Innovation hosts medical device entrepreneur Dr. Christopher Cheng as he presents “From Vascular Motion to Medical Device Innovation, and a Few Tips for Entrepreneurs.” Registration+details: duke.qualtrics.com/jfe/form/SV_...

Friday Roundup: astronauts' wearables report health data to Earth for study; helping the machine to forget; “cognitive surrender” in LLM users; rural-urban gap may spell trouble for clinical AI implementation; when clinicians evade institutional AI governance; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: “accountability pingpong” & health AI; chatbot flattery threatens empathy; using AI scribes at 5 health systems associated with “modest” improvements in EHR time; frontier models confidently describe medical images without actually seeing them; more: aihealth.duke.edu/ai-health-ro...

👇 @dukeaievalgov.bsky.social & @dukemargolis.bsky.social show how health systems can build #AI safety into daily practice through clear oversight, ongoing monitoring, & stronger ways to identify & address issues, so AI supports safe, effective patient care. healthpolicy.duke.edu/publications...

Friday Roundup: docs, AI can't reliably detect deepfake radiology; genetic errors explain cloning failures; gut-brain paper under scrutiny; genetic switch explains big brains; are non-healthcare companies accessing protected data?; health AI and patient trust; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: LLM use may subtly affect viewpoints; study critiques explainable AI; PhD students confront complexities of AI use in academia; new sleeping sickness treatment shows curative potential; concerns grow over RAM shortage as AI gobbles resources; more: aihealth.duke.edu/ai-health-ro...

Join our next Computational and Digital Health Seminar with Dan Ma on Quantitative MRI and Physics-informed Pan-Contrast AI. Register here: duke.qualtrics.com/jfe/form/SV_...

Friday Roundup: what is and is not a stochastic parrot; using LLMs to ID timing, nature of events from chart notes; AI model predicts outcomes from genomic changes; study charts changes from revised kidney function algo; agentic AI for disease surveillance; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: validation study for updated Epic sepsis tool; HealthGPT offers mixed results in diagnosing problems; mathematical model for incentive structures threatening peer review; years of fictionalized clinical vignettes never ID'd as such by journal; more: aihealth.duke.edu/ai-health-ro...

A new perspective in JAMIA by Duke's Michael Cary, Nicoleta Economou, and colleagues examines the National Academy of Medicine’s AI Code of Conduct. The paper highlights a key gap in health AI governance: the lack of explicit nursing representation. academic.oup.com/jamia/advance-

Happening March 24: @dukeaihealth.bsky.social and @dukemargolis.bsky.social are hosting a webinar titled, 'From Deployment to Oversight: Strengthening AI Risk Management and Patient Safety in Health Care,' that will explore strategies to protect patients in the age of AI! To register: duke.is/5/cg7u

Friday Roundup: agentic AI & lab automation; revealing values underlying AI models; neural net classifies animal behavior; pilot program for autonomous clinical care; role of individual states in supporting research; Elsevier unveils lit-searching AI tool; more: aihealth.duke.edu/ai-health-ro...

“Even if you are skeptical about AI’s promise or agnostic about its relevance, the way others use it will shape higher education. Slowing down is not about becoming an AI advocate. It is about developing habits that protect judgment, highlight blind spots and reduce harm.”

Friday Roundup: agentic AI goes rogue; evolutionary link between prokaryotes & eukaryotes; comp sci swamped by AI output; people find AI persuasive, even when they know it’s AI; guidelines for AI usefulness; semantic ablation & the problem with AI writing; more: aihealth.duke.edu/ai-health-ro...

Sudden cardiac arrest remains one of the leading causes of death in the U.S., but we can improve survival. On Feb. 25, thousands of people across NC will learn hands‑only CPR while helping us test how community preparedness can change outcomes. Find a location near you: https://duke.is/5/p925

Friday Roundup: juggling increased work with AI incurs human costs; chatbot responses to health questions affected by phrasing; statin study: no causal link to commonly reported adverse events; evidence quality for edu AI lacking; digital twin for prosthetics; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: evidence on benefit for LLM-human collabs still murky; probing associations between AI use, depressive symptoms; RCT examines human+AI for mammogram reading; geographic AI analysis reveals gaps in MMR coverage; don't abdicate peer review to AI; more: aihealth.duke.edu/ai-health-ro...

The largest randomized trial of medical A.I. —Over 100,000 women in Sweden —radiologist + AI vs 2 radiologists, in follow-up —AI added led to 29% more cancer detected, 44% reduced workload, and —Less cancer dx in subsequent 2 years, and, when found, less aggressive thelancet.com/journals/lan...

Friday Roundup: AI consumes AI, converges on generic representations; AlphaGenome debuts; heritability of human lifespan underestimated; measles spread in SC; gaming or gambling disorder?; disclosure of AI use in journal articles; restoring public trust in science; aihealth.duke.edu/ai-health-ro...

Friday Roundup: going all-in on world models for AI; cows discover tools; challenges to peer review integrity grow; should we get comfortable with imperfect health AI?; too many elderly patients getting meds affecting CNS; questioning AI productivity boosts; more: aihealth.duke.edu/ai-health-ro...

Have a piece from 2025 you are especially proud of? The National Association of Science Writers welcomes entries for our Science in Society Journalism Awards and Excellence in Institutional Writing Awards. Membership not required. Submit by February 1, 2026:

Friday Roundup: AI may be imposing tunnel vision on science; report shows stubborn problems in heart health; taking Health GPT for a test drive; the genetics shaping a dog’s ears; housing, food insecurity among healthcare workers; AI vs clinician deskilling; more: aihealth.duke.edu/ai-health-ro...

A major milestone @ NC Children’s: a $25M transformational gift, the 1st major philanthropic investment supporting our new #behavioralhealth center in Apex. Read more about today’s announcement: duke.is/v/umkg @dukemedschool.bsky.social | @dukechildrens.bsky.social | @unchealth.bsky.social (1/2)

Monica Bertagnolli — former NIH director and surgical oncologist — has been elected President of the National Academy of Medicine. Bertagnolli has championed efforts to advance #HealthInnovation and #HealthEquity, and will be the first woman to lead the NAM. Learn more: https://ow.ly/vPhb50XXuw4

After studying the world’s most active communities, Duke evolutionary anthropologist Herman Pontzer, PhD, found something shocking: we all burn about the same amount of calories. #AcademicSky #MedSky https://medschool.duke.edu/stories/why-exercise-doesnt-burn-more-calories-and-why-thats-not-point

My department, Duke Neurobiology, is searching for a new chair. Ad below. Come work with me, @jmgrohneuro.bsky.social @ennatsew.bsky.social @jorggrandl.bsky.social @jnklab.bsky.social @sbilbo.bsky.social @neurocircuits.bsky.social and many other amazing folks! @dukemedschool.bsky.social

Dr. Nicholas Christakis (@nachristakis.bsky.social ) will be joining us as a Keynote Speaker at the Conference on Society-Centered AI 2026 Feb 12-14th! Learn more and Register here: sites.duke.edu/scai/

Friday Roundup: pinball-like LLM parameters; off-target vaccine benefits; humility & curiosity for LLMs; OpenAI unveils ChatGPT Health; international alternatives to PubMed; relationship between flu & heart attack; LLMs still at risk from prompt injection; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: choosing measures for AI performance; problems with AI simulation in behavioral studies; AI ethics text full of bogus citations; testing chatbots on subspecialty boards; looking back on computer science in 2025; ML spots liver disease in EKGs; more: aihealth.duke.edu/ai-health-ro...

Our guidance regarding performance measures for medical AI models is finally out! - Stop bashing AUROC, although it does not settle things - Calibration and clinical utility are key - Show risk distributions - Classification statistics (e.g. F1) are improper www.thelancet.com/journals/lan...

Friday Roundup: ambient scribing vs clinician burnout; merits of explainable AI in healthcare; PFAS & infant health; tattoos & immune response; problematic data leads to scores of retractions; AI for manuscript review; autoencoders for high-dimensional data; more: aihealth.duke.edu/ai-health-ro...

Two DIBS Network Faculty elected to National Academy of Inventors! Dr. Adam Wax: methods for imaging retinas & tracking biomarkers of Alzheimer's disease. Dr. Arovind Asokan: viral methods for gene therapies incl. for the brain. @dukeubme.bsky.social @dukesurgery.bsky.social tinyurl.com/DIIBS-NAI

Friday Roundup: considering healthcare AI bubbles; Cochrane reports on HPV vaccine & cervical cancer; AI slop crops up in autism paper; academic libraries in the AI age; lonely people turn to chatbots; AI-written peer reviews & papers show up at conference; more: aihealth.duke.edu/ai-health-ro...

Institute Director Mark McClellan and Research Director Christina Silcox have released a statement on the new CMS ACCESS Model, which aims to improve the health of people with chronic disease and covered by Traditional Medicare, through technology-supported care. Read our statement: duke.is/j/tqrq

What started as a fight against one virus led to breakthroughs in immunology, cancer treatment, hepatitis care, pandemic preparedness, and more. Read about 10 major advances that owe their roots to HIV research ➡️ https://medschool.duke.edu/news/10-groundbreaking-advances-grew-out-hiv-research

Friday Roundup: practice licenses for AI?; drones deliver defibrillators fast; continuous evaluation of predictive models; light pollution messes with ecosystem metabolism; AI competencies for docs; cancer vax trial shows encouraging results 2 decades later; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: puzzle-solving illuminates new paths for AI; link between Epstein-Barr & lupus; frameworks for reporting GAI research; new chip extends quantum computing; creative potential of AI; updated Turing Test spots key differences between human, AI; more: aihealth.duke.edu/ai-health-ro...

NEW PAPER: "Reporting guidelines for studies involving generative artificial intelligence applications: what do I use, and when?" --> www.nature.com/articles/s41... #HealthTech #ClinicalAI #MachineLearning #MedAI #AIinMedicine #TransparentAI #HealthInnovation #GenAI

Friday Roundup: offloading moral agency to chatbots; benefits & limits of RAG-based LLMs for medical teaching; tiny tyrannosaur confirmed; bogus AI-penned letters swamping journals; LLMs for data extraction; AIs struggle to parse out-of-sequence comic strips; more: aihealth.duke.edu/ai-health-ro...

Friday Roundup: using LLMs for research changes quality of downstream product; how activity occurs modulates exercise benefits; script sleuths for bogus citations; study finds bots distort news; humans and LLMs apply different priorities when asked to judge; more: aihealth.duke.edu/ai-health-ro...

A recent study led by DCRI’s Jennifer Rymer suggests one long-held DST belief, that these time jumps can lead to an uptick in myocardial infarctions, may not be the case. Read more: www.medscape.com/viewarticle/...

Duke research saves lives. As funding landscapes shift, the impact of Duke’s research remains clear: it’s personal, it’s powerful, and it’s essential. 🔗 Read how Duke research continues to make a difference: https://today.duke.edu/2025/10/saved-duke-research #AcademicSky #MedSky

Friday Roundup: mimicking brain modularity for better AI models; bumblebee empathy; “fragility” of GPT-5 output in medicine; mRNA COVID vaccines boost immune response vs cancers; human vs LLM in diagnosis; evaluating memorization risk in EHR-trained models; more: aihealth.duke.edu/ai-health-ro...