[Large language models in rheumatology : New ways of knowledge transfer]

Medical knowledge is growing exponentially in rheumatology, posing increasing challenges for knowledge dissemination among physicians, educators and patients. Traditional information and learning formats are reaching their limits in view of the rapid emergence of new clinical studies, guidelines and treatment concepts. Large language models (LLMs) offer the possibility to structure, synthesize and contextually adapt extensive and complex information within a short time. This opens new perspectives for clinical decision support, medical education, patient education and scientific work in rheumatology. This article systematically categorizes the use of LLMs across these fields of application and discusses the opportunities, risks and practical implications using selected examples that colleagues can immediately apply. Early data suggest a high potential for use and growing acceptance among physicians, students and patients; however, significant challenges remain. These include concerns regarding the validity and transparency of the generated content, potential biases, data protection issues and the risk of uncritical adoption of artificial intelligence (AI)-generated recommendations. The use of LLMs can make rheumatological knowledge rapidly, individually, and easily accessible but does not replace medical responsibility or clinical expertise. An evidence-based evaluation, clear regulatory framework conditions, safe integration into existing workflows and targeted training and governance concepts are essential to harness the potential of LLMs for knowledge dissemination in rheumatology in a sustainable and responsible manner.