• Background

  Large language models (LLMs) are increasingly being integrated into neurosurgical practice, education, and research, demonstrating significant potential and challenges.

• Objective

  To raise the importance of transparent, collaborative, and open-source AI development and literacy.

• Methods

  A letter to the editor discussing the use of retrieval-augmented generation (RAG) in AtlasGPT, emphasizing the complexities of LLM, the need for explainable AI (XAI), and proposing open-source frameworks like Neura for improved user engagement and representativeness.

• Project type: editorial

• Team status: complete

• Initiator: Sami Barrit, MD

• Co-investigators: Salim El Hadwe, MD, MSc; Romain Carron, MD, PhD, Joseph R Madsen, MD

• Project status: done