Rachel Hendrick writes: “The limitations of generative artificial intelligence (AI) are myriad: They don’t provide citations, they are prone to hallucinations, there is no way to reproduce results, and there are major issues with copyright and user privacy. While retrieval augmented generation (RAG), a framework that creates an application for generative AI large language models, doesn’t solve all these problems, it begins to address the concerns of the academic community. At its core, RAG is the difference between generative AI for fun and generative AI as a legitimate research tool.”