Duke Health AI Evaluation & Governance Program

Duke Health AI Evaluation & Governance Program Director Nicoleta Economou, PhD, will deliver the keynote at the upcoming SAS Health Care Community Connection virtual event, which is open to the global health care community. Register here: sas.com/en_us/events/2

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

Duke's Dr. Eric Poon, will represent @dukeaievalgov.bsky.social's work at the upcoming #HIMSS2026. The E&G Program has collaborated with Avanade to develop SAIGE (Smart AI Governance Engine), an important step toward advancing responsible health AI. healthaigovernance.duke.edu/abcds-oversi...

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-

Our article, “...Consensus-Based Considerations for AI Usefulness in Healthcare,” is now published in The American Journal of Bioethics. @dukeaievalgov.bsky.social is excited to collaborate with colleagues across CHAI's Usefulness Workgroup for this publication: tandfonline.com/doi/full/10.10

Duke Health AI E&G at the Duke Department of Medicine AI Symposium Duke Department of Medicine AI Symposium (January 30, 2026): In January, ABCDS Oversight Assistant Director Sophia Bessias joined Dr. Clay Musser and Dr. Misha Tran at the Department of Medicine AI Sy...

ABCDS Oversight Assistant Director Sophia Bessias recently joined Dr. Clay Musser and Dr. Misha Tran at the Duke Department of Medicine AI Symposium for a panel discussion of clinical AI and evaluation, moderated by Dr. Juliessa Pavon. Learn more here: healthaigovernance.duke.edu/news/duke-he...

We congratulate Erasmus MC's Michel van Genderen, MD, PhD, on his appointment as Adjunct Associate Professor at Duke. Dr. van Genderen will partner with @dukeaievalgov.bsky.social to expand opportunities for joint research and to develop frameworks for trustworthy, evidence-based #healthAI.

Using real #COVID19 data, the study found that hospitals can learn faster, reduce bias, and make safer treatment decisions using fewer patient data. The work highlights a practical path for bringing responsible #AI into everyday clinical care.

A federated learning framework for ethical dynamic treatment allocation across heterogeneous hospitals In this paper, we propose an adaptive federated learning framework to learn optimal treatments for individual hospitals that possibly serve different …

New research published in the Journal of Biomedical Informatics by Duke researchers shows how hospitals can use #AI to learn collaboratively which treatments work best while keeping patient data private. www.sciencedirect.com/science/arti...

As 2025 comes to a close, we remain focused on what matters most: our patients. By advancing safe, fair, and high-quality algorithmic solutions, we’re working to improve and responsibly transform healthcare. Thank you for being part of this progress. Wishing you Happy Holidays! 🥳

Today at 2:00 PM EST: @dukeaievalgov.bsky.social Director, Nicoleta Economou, PhD, will join Avanade's David Galich and David Gormley for the final installment of their Smart AI Masterclass series, where the speakers will discuss #AgenticAI. Register now. events.teams.microsoft.com/event/ebdde5...

👇 @dukeaievalgov.bsky.social welcomes Elijah Richardson, MPH, who is the new Evaluation & Governance Lead for our flagship ABCDS Oversight initiative. Learn more about his new role and his professional background here: healthaigovernance.duke.edu/personnel/el...

The appointment highlights growing recognition of @dukeaievalgov.bsky.social program's work in responsible #AI oversight, advancing ethical, sustainable, and human-centered innovation that keeps quality and accountability at the forefront. healthaigovernance.duke.edu

AI at Duke Initiative Strengthens Leadership and Expands Faculty Engagement - AI at Duke The AI at Duke steering committee continues to gain momentum as it enters a critical phase of strategic planning and university-wide engagement. Launched by Provost Alec Gallimore in May, the initiati...

Nicoleta Economou, PhD, director of @dukeaievalgov.bsky.social has been appointed to the AI at Duke Steering Committee and will also serve as liaison for the Trustworthy & Responsible AI Advisory Committee. Provost Alec Gallimore launched the AI at Duke initiative in May 2025. duke.is/g/newp

Aligned with Duke Health's broader mission, the @dukeaievalgov.bsky.social is advancing a streamlined approach to AI evaluation & governance, establishing rigorous frameworks for responsible oversight and fostering a culture of best practices across the health system. healthaigovernance.duke.edu

The @dukeaievalgov.bsky.social team was excited to present our work at the Duke Summit on AI for Health Innovation. Throughout the event, we observed meaningful engagement and genuine curiosity from attendees regarding our program, the ABCDS Oversight Initiative, and our published research.

From Chaos to Coordination: Reimagining Healthcare Emergency Management with AI This event is a part of the AHEM Expert Panel Series Join us for the fifth session in the Academic Healthcare Emergency Management (AHEM) Expert Panel Series, where leading voices in AI, healthcare…

@dukeaievalgov.bsky.social Director Nicoleta Economou, PhD, recently joined leading experts in #AI, data science, public health, and healthcare emergency management for an AI Panel Series hosted by the Academic Healthcare Emergency Management (AHEM) Consortium. ahemconsortium.org/from-chaos-t...

👇 @dukeaievalgov.bsky.social Director Nicoleta Economou, PhD, recently delivered a keynote at the National Evaluation Center of Quality and Technology in Health Conference in Athens, which highlighted @dukehealth.bsky.social’s leadership in advancing trustworthy and responsible #AI for healthcare.

Agentic AI, Generative AI and AI Governance Explore how agentic AI, generative AI and AI governance are driving transformation across health and life sciences.

👇 @dukeaievalgov.bsky.social Director Nicoleta Economou, PhD, joined other AI leaders on a @sassoftware.bsky.social-hosted webinar exploring how agentic AI, generative AI, and AI governance are reshaping health and life sciences. Watch the webinar on demand here: www.sas.com/en_us/webina...

Google’s healthcare AI made up a body part — what happens when doctors don’t notice? The basilar ganglia does not exist.

🤔Typo or something more? The Verge's Hayden Field spoke with @dukehealth.bsky.social's Chief Data Scientist Michael Pencina, PhD, about the risks of "hallucinations" in medical #LLMs like Google's Med-Gemini, which mistakenly cited an anatomical structure that doesn’t exist, in a recent preprint.

Whether you are an informaticist, clinician, policy nerd, or governance geek, this curated collection of papers offers practical frameworks, regulatory foresight, and real-world applications to inform and inspire your work.
healthaigovernance.duke.edu/our-work/pub...

Fall Reading on AI Governance Looking to add some academic heat to your fall reading list? The Duke Health AI Evaluation & Governance Program is spotlighting our latest research on algorithm oversight, LLM evaluation, fairness, an...

Looking to add some academic heat to your fall reading list, the @dukeaievalgov.bsky.social is spotlighting our latest research on algorithm oversight, #LLM evaluation, fairness, and safety in #healthAI. healthaigovernance.duke.edu/news/fall-re...

Duke innovates on implementing and assessing AI in health spaces Researchers are creating evaluation frameworks to ensure the safe usage of LLMs, like ChatGPT, at Duke Health

North Carolina Public Radio (@wunc.org) reports on @dukehealth.bsky.social's national leadership in health #AI governance, spotlighting two new evaluation frameworks developed by Duke researchers to assess the safety, reliability, and clinical performance of #LLMs.
www.wunc.org/health/2025-...

Duke researchers want to ensure AI is safe in a health care setting Health systems across the country are investing heavily in AI as a way to reduce burnout for providers and potentially improve care

👇@axios.com reports on how a team of Duke researchers led by Chief Data Scientist Michael Pencina has developed two important evaluation frameworks to assess the safety, accuracy, and reliability of large language models (LLMs) in clinical care. axios.com/local/raleig... @dukehealth.bsky.social

An evaluation framework for ambient digital scribing tools in clinical applications npj Digital Medicine - An evaluation framework for ambient digital scribing tools in clinical applications

These frameworks offer a new approach to ensuring that #AI systems used in clinical settings meet the highest standards of quality and accountability:

Read the NPJ Digital Medicine paper here: rdcu.be/etzTt
Read the JAMIA paper here: academic.oup.com/jamia/articl... 2/2

Duke University Sets New National Standard for Safe, Scalable AI in Health Care DURHAM, N.C. — Duke University School of Medicine researchers have developed two pioneering frameworks designed to evaluate the performance, safety, and reliability of large-language models in health ...

Duke researchers, led by Dr. Michael Pencina and Dr. Chuan Hong, have developed two pioneering frameworks designed to evaluate the performance, safety, and reliability of large language models (LLMs) in healthcare. corporate.dukehealth.org/news/duke-un... 1/2

Bridging the Gap Between Developers and Implementers in Health AI This Viewpoint describes opportunities for health care organizations to close gaps between the development and implementation of artificial intelligence using third-party enablers for training, adhere...

New in @jamahealthforum.bsky.social: Duke’s Michael Pencina and colleagues make a case for third-party enablers, smarter local governance, and shared evidence networks to make health #AI safe, scalable, and truly impactful. jamanetwork.com/journals/jam... @dukehealth.bsky.social

Welcome! You are invited to join a meeting: Shaping the Future of Responsible Health AI: Introducing the Duke Health AI Evaluation & Governance Program. After registering, you will receive a confirmat... Speakers: Michael J. Pencina, PhD, Chief Data Scientist, Duke Health Nicoleta J. Economou-Zavlanos, PhD, Director, Duke Health AI Evaluation & Governance (E&G) Program Sophia Bessias, MPH, MSA, Assist...

ATTN: Duke faculty, staff, and students: Join @dukeaievalgov.bsky.social for a special webinar: 'Shaping the Future of Responsible Health #AI'. Learn how Duke is advancing the deployment of trustworthy and responsible AI across the institution. Register: qrcc.me/swb9tiwh8cw6... (Duke login required)

This is just the beginning! @DukeAIEvalGov welcomes collaborations and partnerships to shape an #AI-powered future that prioritizes trust, transparency, and transformative impact. Let’s build it together! Contact us: healthaigovernance.duke.edu/contact-us 2/2

A website launch flyer showcasing the front page of the new Duke Health AI Evaluation & Governance website on various screens.

Exciting News! @dukeaievalgov.bsky.social Program proudly launches its new website. The Program promotes the safe, fair, and effective use of #AI in healthcare, ensuring AI innovations improve patient care ethically. Visit & explore here: healthaigovernance.duke.edu 1/2