Researchers propose a dual strategy—combining retrieval-augmented generation with publisher-built academic LLMs—to combat AI's dangerous tendency to fabricate citations in medical literature. This response to critical feedback underscores the non-negotiable need for accuracy where flawed references could impact patient care. The path forward demands unprecedented collaboration between AI developers, publishers, and human experts to safeguard scientific integrity.