reviewing-ai-output-groundedness

Human review protocol for whether AI output is grounded in its sources: claim-by-claim attribution, hallucination classes, citation verification, sampling that is defensible, and a recorded attestation because groundedness cannot be fully automated. Also covers the responsible-AI pass — harmful output, bias, disclosure and privacy. Use when reviewing a RAG or summarisation feature, when factuality matters more than…