SCCMPod-577 CCE: Exploring AI Chatbots in the PICU

Generative AI is rapidly reshaping healthcare, creating new opportunities and raising new questions about how to use the technology effectively and ethically. In this episode of the Society of Critical Care Medicine (SCCM) Podcast, host Maureen A. Madden, DNP, RN, CPNP-AC, CCRN, FCCM, speaks with Brandon Hunter, MD, FAAP, about his article, “Feasibility of a Large Language Model Chatbot to Support Parental Understanding in the PICU,” published in the April 2026 issue of Critical Care Explorations. Dr. Hunter explores the potential role of large language model (LLM) chatbots in helping families navigate the complexity of pediatric critical illness. The study evaluated a HIPAA-compliant chatbot that was connected to selected electronic health record data and allowed parents to ask questions about their child’s condition, treatments, laboratory results, and prognosis. The results were promising, with the chatbot answering most questions accurately and returning personalized responses shaped by available clinical information. The podcast discussion examines both the opportunities and challenges of AI-powered communication tools. Dr. Hunter explains how modern LLMs generate responses; highlights the importance of validating chatbot outputs; and describes hazards such as hallucinations, bias, and sycophancy, in which models may overly agree with users rather than prioritize accuracy. The conversation further explores the growing use of AI across healthcare; the need for rigorous governance and evaluation frameworks; and the importance of ensuring that technological innovation serves patients, families, and clinicians safely and equitably. This episode offers a thoughtful examination of one of the most rapidly evolving areas in healthcare, highlighting both the opportunities and responsibilities that accompany the integration of AI into critical care practice. Resources referenced in this episode: Hunter RB, Thammasitboon S, Rahman S, et al. Feasibility of a large language model chatbot to support parental understanding in the PICU. Crit Care Explor. 2026;8(4):e1378. Omar M, Soffer S, Agbareia R, et al. Sociodemographic biases in medical decision making by large language models. Nat Med. 2025;31:1873–1881.