KPMG's AI Governance Report Undermined by AI Hallucinations

GPTZero researchers identified 40 fabricated or incorrect citations in a KPMG report examining business AI deployment, retaining only 5 accurate references out of 45. The hallucinations — plausible-sounding but nonexistent author names, publication dates, and journal titles — bear the signature of large language model errors. The irony compounds the credibility damage: a document intended to shape client thinking about AI governance was itself compromised by the failure mode it could have been expected to address.