If you mean the AI help assistant, it's just wrong (hallucinating; it tends to either give good answers or nonsense answers, depending on whether the RAG system finds relevant docs or it's something it can infer from the base model).
Statistics: Posted by Antony Lewis — February 23 2025