Enhancing Rare Disease Education through AI-Driven Podcast Generation.

ObjectiveRare diseases, including many rare genetic epilepsies and neurodevelopmental disorders, present significant challenges in timely diagnosis, treatment, and patient education due to their rare incidence, complex clinical nature and lack of standardized care pathways. Despite advancements in genetic testing, knowledge dissemination remains inadequate, contributing to delayed diagnosis and inconsistent management. Addressing these gaps requires innovative educational approaches tailored to diverse audiences, including patients, caregivers, and non-specialist healthcare providers. MethodsWe explored the potential of AI-driven podcast generation as a scalable solution for rare disease education. We designed a tutorial for the community to use Googles NotebookLM, a tool powered by Large Language Models and text-to-speech technology to generate podcasts. ResultsEight examples were created from research papers on rare epilepsies (SCN2A-, CACNA1A-, SYNGAP1-, chromosome 8p and SLC6A1-related disorders) and a complex genetic research topic (epilepsy polygenic risk scores) in English, Spanish or German. The AI-generated podcasts featured conversational overviews delivered by virtual hosts, with customizable style and tone enabling personalized content creation for different audiences. Feedback from 20 stakeholders, including patient advocacy leaders, researchers, and clinicians, highlighted strong enthusiasm for this approach, particularly for under-resourced patient communities. Respondents praised the accessibility, quality of language translation and educational value of the podcasts, noting their potential to bridge the gap between complex research findings and practical patient care. Key recommendations for improvement included ensuring scientific accuracy through expert review, enhancing clinical depth, reducing redundancy, and incorporating structured episode elements such as introductions and summaries. SignificanceThe study underscores how AI-driven podcasting can democratize access to high-quality medical information, offering a flexible, multilingual platform that adapts to the needs of global rare disease communities. By refining this approach to include greater oversight and targeted content development, AI-generated educational podcasts could play a pivotal role in rare disease knowledge dissemination, ultimately improving patient outcomes and empowering stakeholders across healthcare systems. Key pointsO_LIAI-driven podcasts bridge knowledge gaps in rare disease education via scalable, multilingual, and accessible formats. C_LIO_LIOur method enables rapid podcast generation, offers tailored content for diverse audiences in various languages with fair clinical depth. C_LIO_LIAI podcasts reduce barriers for caregivers, patients, and clinicians, democratizing access to medical knowledge. C_LIO_LIHuman oversight and expert supervision are required and ensures scientific accuracy, balancing accessibility with professional rigor for impactful education. C_LI