BIO: As CEO of Seattle Children’s, Dr. Christopher Longhurst, MD, MS leads a world-class team of healthcare providers, researchers, and staff who are united by Children’s mission to provide hope, care and cures to help every child live the healthiest and most fulfilling life possible.
A practicing pediatrician for the past 25 years, Dr. Longhurst joined Seattle Children’s in 2026 after serving in a dual role as chief medical officer (CMO) and chief digital officer (CDO) at UC San Diego Health, as well as Professor of Biomedical Informatics and Pediatrics at UC San Diego School of Medicine. Combining a passion for innovation with the drive to improve quality, safety, equity and patient experience, Longhurst was instrumental in securing a philanthropic gift to establish the Joan & Irwin Jacobs Center for Health Innovation where he served as the center’s founding executive director leading the AI portfolio across the system.
Dr. Longhurst now serves as an affiliate Professor of Pediatrics and Biomedical Informatics at the University of Washington and continues to contribute thought leadership and scholarship in care quality, patient safety and health informatics. He has published over 150 peer-reviewed articles in journals like the New England Journal of Medicine, JAMA and Pediatrics, and serves on the National Academy of Medicine steering committee for Patient Safety in the AI Era.
Before joining UC San Diego Health, Dr. Longhurst spent 15 years at Stanford University, serving as chief medical information officer for Stanford Children’s Health, where he led efforts to improve children’s health and provider workflow using information technology. He also founded and led the nation’s first accredited clinical informatics fellowship at Stanford, where he was a clinical professor of pediatrics and biomedical informatics.
Dr. Longhurst is an elected fellow of the prestigious American College of Medical Informatics. He earned his medical degree and Master of Science in Medical Informatics from UC Davis, completed his residency at Stanford University, and holds a Bachelor of Science from UC San Diego.
ABSTRACT: Digital health is rapidly expanding due to surging healthcare costs, deteriorating health outcomes, and the growing prevalence and accessibility of mobile health and wearable technologies. Recent technological advancements make it possible to closely and continuously monitor individuals using multiple measurement modalities in real time. We are collecting and integrating such wearables data with clinical information to gain a more precise understanding of health and disease and develop actionable, predictive health models for improving outcomes. We are simultaneously developing open source data science and machine learning tools for the digital health community, including the Digital Biomarker Discovery Pipeline (DBDP), to facilitate the use of mobile device data in healthcare.
BIO: Jessilyn Dunn, PhD, is Associate Professor of Biomedical Engineering and Biostatistics & Bioinformatics at Duke University. She directs the BIG IDEAs Lab, which is focused on digital health innovation, wearable sensors, and the development and validation of AI-driven digital biomarkers. Dr. Dunn is the Principal Investigator of research initiatives funded by the NIH, NSF, and FDA which are developing digital biomarkers of conditions ranging from pre- and type 2 diabetes to influenza-like illness to Opioid Use Disorder. She sits on the Google Consumer Health Advisory Panel and is a recipient of the NSF CAREER Award and the IEEE EMBS Early Career Achievement Award for her leadership and innovation across engineering and medicine.
ABSTRACT: This talk will introduce Dr. CaBot (https://arxiv.org/abs/2509.12194), an agentic AI system that my lab built in collaboration with physicians to emulate an expert diagnostician by generating written and slide-based presentations based on a case presentation alone. I will discuss several cases, including a recent case in which CaBot became the first AI system to generate a diagnosis published in the 100+ year history of the NEJM CPCs (https://www.nejm.org/doi/abs/10.1056/NEJMcpc2412539) and a public demonstration of CaBot covered recently in the New Yorker (https://www.newyorker.com/magazine/2025/09/29/if-ai-can-diagnose-patients-what-are-doctors-for).
BIO: Arjun (Raj) Manrai, PhD is an Assistant Professor in the Department of Biomedical Informatics at Harvard Medical School, where he leads a research lab that works broadly on applying machine learning and statistical modeling to improve medical decision-making. Raj is also a founding Senior Deputy Editor of NEJM AI and co-host of the NEJM AI Grand Rounds podcast. Focus areas for Raj’s research group include artificial intelligence in diagnostic and management reasoning, evaluating and improving common clinical equations, cardiovascular disease and kidney disease, decision making across populations, and reproducibility and safety challenges for medical artificial intelligence. His work has informed national and international clinical practice guidelines that affect millions, and his lab recently created Dr. CaBot, which generated the first AI diagnosis published in the 100+ year history of the NEJM Clinicopathological Conferences. His work has been published in the New England Journal of Medicine, JAMA, and Science, presented at the National Academy of Sciences, and featured in the New York Times, New Yorker, Wall Street Journal, and NPR.
ABSTRACT: Evaluations are crucial for shaping model development, both directly through training and indirectly
by influencing what model developers focus on. This is especially true in medical domains where evaluations are complex. We will briefly discuss PanCanBench, a benchmark for pancreatic cancer questions. Then, we discuss our work on training DR Tulu, a long-form deep research model, by using RL with rubrics that evolve as the model trains, thereby providing discriminative training signals even on complex deep research tasks.
BIO: Pang Wei Koh is an assistant professor in the Allen School of Computer Science and Engineering at the University of Washington. His research interests are in the theory and practice of building reliable machine learning systems. His research has been published in Nature and Cell, featured in The New York Times and The Washington Post, and recognized by the AI2050 Early Career Fellowship, MIT Tech Review Innovators Under 35 Asia Pacific Award, Google ML and Systems Junior Faculty Award, a Singapore AI Visiting Professorship, and best paper awards at ICML, KDD, and ACL. He received his PhD and BS in Computer Science from Stanford University. Prior to his PhD, he was the 3rd employee and Director of Partnerships at Coursera.
ABSTRACT: As generative AI steadily integrates into the social fabric of daily life, therapy and companionship have emerged as dominant use cases. An estimated 1 in 4 U.S. adults have utilized a large language model for mental health support, often operating entirely outside formal clinical oversight or
supervised healthcare systems. However, standard machine learning safety benchmarks are
fundamentally ill-equipped for this domain. By optimizing for nonmaleficence through single-turn
text classification or token-filtering, existing frameworks miss the interactional, accumulative
dynamics—such as dependency formation, therapeutic ruptures, and the reinforcement of cognitive
distortions—that dictate real-world psychological harm over time. This talk introduces a new
community-wide effort to establish the evaluation frameworks, multi-stakeholder governance, and
technical infrastructure required to ensure these systems operate with clinical integrity. At the center
of this effort is the open-source Virtual Sandbox: a continuous evaluation environment that
simulates multi-turn, multi-session interactions at scale. By leveraging a comprehensive attribute
taxonomy and a stateful simulation engine, the Sandbox generates psychologically plausible
synthetic users whose internal emotional states dynamically evolve in response to model behavior.
Rather than producing static, gamifiable leaderboard scores, this architecture yields
multi-dimensional Safety Profiles that explain exactly why a model succeeds or fails, for which
populations, and through what psychological mechanisms. Ultimately, we demonstrate how this
continuous infrastructure translates clinical principles into enforceable market signals, bridging the
gap between cutting-edge AI development and safe, beneficial human-centered AI applications.
BIO: Tim Althoff is the Jean-Loup Baer Associate Professor at the University of Washington’s Allen School. Tim’s research seeks to better understand and empower people through data and computation. His AI research has directly improved mental health services utilized by over ten million people and informed federal policy. A Stanford Ph.D. and NSF CAREER Award recipient, his work has earned numerous Best Paper Awards (including WWW, ACL, ICWSM and UbiComp) and the 2019 SIGKDD Dissertation Award. His research is frequently featured in major global outlets such as The New York Times, The Economist, and The Wall Street Journal.
ABSTRACT: Large language models (LLMs) are increasingly used for mental health support, but existing AI alignment methods typically optimize therapeutic objectives in isolation, leading to trade-offs between empathy, safety, and patient autonomy. We introduce a multi-objective direct preference optimization (MODPO) framework that jointly aligns LLMs across multiple clinically grounded therapeutic criteria using patient-derived preference data. Through large scale persona-based evaluation and blinded clinician validation, we show that multi-objective alignment produces responses that are consistently preferred over baselines while maintaining non-negotiable safety standards.
BIO: Saadia Gabriel is an Assistant Professor of Computer Science at UCLA, where she leads the Misinformation, AI and Responsible Society Lab. Her work aims is to develop NLP technologies that improve diverse users’ well-being, critical thinking skills, and civic agency without displacing human autonomy. Her research has received several best paper nominations or awards, and has been covered by a wide range of media outlets like Forbes and TechCrunch. She was named on Forbes’ 30 under 30 2024 list and has received research awards from Google and Amazon. She previously was a NYU Data Science Faculty Fellow and MIT CSAIL Postdoctoral Fellow. She received her PhD from the University of Washington.
ABSTRACT: As AI systems are introduced into high-stakes domains, decisions around their use and governance
must center the specific contexts and communities they affect. My research explores participatory approaches that support community control and local agency in AI design and governance. I’ll first discuss challenges to incorporating meaningful participation into general-purpose foundation models, highlighting a tension between local agency and scale, and proposing more domain-specific opportunities for participatory governance. Building on this, I’ll share findings from an 18-month ethnographic collaboration with mental health practitioners evaluating LLM counselors agains professional codes of conduct. We surface violations that persist across model architectures and prompt strategies, such as deceptive empathy, where apparent emotional responses create false trust with vulnerable users. I’ll end with open questions around what sustained, community-led governance of AI in mental health should look like: how can practitioners collectively articulate clear ethical guidelines and boundaries, and how do we close the gap between those guidelines and actual, enforceable accountability?
BIO: Dr. Harini Suresh is an Assistant Professor of Computer Science at Brown University, a core faculty in the Center for Technological Responsibility, Reimagination, and Redesign (CNTR), and a CRA Trustworthy AI Fellow. She leads the Data in Society Collective (DISCO Lab), which studies and builds AI systems that center individual and community agency. Her research takes an interdisciplinary and human-centered approach, building on sustained collaborations with domain experts including journalists, civil society organizations, and mental health practitioners. Before joining Brown, Harini was a postdoctoral researcher at Cornell Tech, examining the limitations of participatory approaches in the era of “general-purpose” AI systems. She completed her PhD, M.Eng. and B.Sc. in Computer Science at MIT.
BIO: Lucy Lu Wang is an Assistant Professor at the University of Washington Information School, where she leads the Language Accessibility Research (LARCH) lab. She holds adjunct appointments in the Paul G. Allen School of Computer Science & Engineering, Department of Biomedical Informatics & Medical Education, and Department of Human Centered Design & Engineering at the University of Washington, and is a Research Scientist at the Allen Institute for AI (Ai2). Her work spans scholarly document understanding, document accessibility, scientific evidence synthesis, and health communication. She focuses on developing language technologies to improve access to and understanding of information in high-expertise domains like science and healthcare, with an emphasis on dataset development and evaluation practices. Her work on supplement interaction detection, document accessibility, and academic publishing trends have been featured in media outlets such as Geekwire, Boing Boing, Axios, VentureBeat, and the New York Times.
ABSTRACT: This talk will share the journey and learnings of the PROSPECT lab, the data science arm of the Zuckerberg San Francisco General Hospital. Our mission is to apply AI/ML and digital technologies to improve health outcomes and equity in vulnerable and underserved populations. We will start from the very beginning—how our resource-constrained, urban safety-net hospital decided to commit funds towards a data science team—and walk through key projects that have shaped our team’s role in the hospital today.
BIO: Jean Feng is an Associate Professor in the Department of Epidemiology and Biostatistics at the University of California, San Francisco and the UCSF-UC Berkeley Joint Program in Computational Precision Health. As a principal investigator at the UCSF-Stanford Center of Excellence in Regulatory Science and Innovation (CERSI), she collaborates closely with researchers from the US Food and Drug Administration to develop methods that improve the safety, reliability, and interpretability of artificial intelligence (AI)/machine learning (ML) algorithms in healthcare. She is also the data science lead on the PROSPECT team, the digital innovation task-force for the Zuckerberg San Francisco General Hospital.
ABSTRACT: This talk shares practical lessons from deploying AI at scale in a large academic medical center. I’ll cover Stanford Healthcare’s responsible AI framework, evaluation and monitoring approaches, and clinical and operational workflows enabled by the ChatEHR platform.
BIO: Nikesh Kotecha is currently the Head of Data Science at Stanford Health Care (SHC). Operating at
the intersection of research and clinical implementation, Nikesh has built and scaled a cross-functional data science group charged with a two-fold mandate: establishing frameworks for organization-wide AI adoption, and leveraging in-house capabilities to solve problems for the enterprise. Previously, Nikesh led informatics at the Parker Institute for Cancer Immunotherapy and co-founded Cytobank Inc. (acquired by Danaher Life Sciences), a cloud-based single-cell analytics company. He also served as an Entrepreneur-in-Residence at FactoryHQ and delivered enterprise solutions at TIBCO Spotfire.
Nikesh holds a Ph.D. in Biomedical Informatics from Stanford University and a B.S. in Bioengineering
from Boston University. He stays involved in entrepreneurship and medicine through his work with StartX and as an Adjunct Professor at Stanford Medicine.
BIO: Dr. Clara Lin is the Chief Medical Information Officer (CMIO) and VP of Digital Health and Informatics at Seattle Children’s, where she leads organizational growth through informatics and innovation. Board certified in Internal Medicine, Pediatrics, and Clinical Informatics, she possesses a unique perspective on the intersection of technology and patient care. In her current role, Dr. Lin facilitates the organization’s clinical and operational AI portfolio and co-chairs the Artificial Intelligence Review Board (AIRB). Her mission is to leverage technology to improve quality of care and clinician wellness. Outside of her leadership at the hospital, she serves as a Clinical Associate Professor in the Department of Pediatrics and Affiliate Associate Professor in the Department of Biomedical Informatics and Medical Education (BIME) at the University of Washington.
