Our approach for truthful foundation model (LLM) question-answering using RAG methods in robotics was accepted by IEEE ICRA.