KPMG Withdraws AI Report Over Hallucinated Data Claims

Global professional services giant KPMG has removed its recent publication, "Redefining excellence in the age of agentic AI," from all its platforms following revelations of widespread inaccuracies. The report, cited as published in October 2025, contained numerous claims about organizations' use of artificial intelligence that were subsequently refuted by several prominent entities. Among those disputing the assertions were financial institution UBS, the UK’s National Health Service, Swiss Federal Railways, and Transport for London, all of whom indicated the report’s statements regarding their AI deployments were either untrue or significantly misleading.

The inconsistencies were initially brought to light by research group GPTZero, which informed the Financial Times that the errors likely stemmed from "AI hallucinations"—a phenomenon where generative artificial intelligence fabricates information. This suggests that KPMG, a firm renowned for its professional insights, may have utilized AI tools to help draft a report on AI itself, inadvertently showcasing the critical need for human oversight. A KPMG spokesperson confirmed the report’s withdrawal, stating that an internal investigation is underway. The firm emphasized its expectation that all personnel strictly adhere to internal guidelines for responsible AI use, which mandate human validation and verification of all independent sources prior to publication. This adherence is crucial to maintaining the integrity and factual accuracy expected from professional content.

This incident underscores a critical challenge for professional services firms integrating AI into their workflows, particularly for research and content creation. The necessity for stringent human review and robust fact-checking mechanisms is becoming increasingly evident to prevent the dissemination of misinformation and protect professional credibility. The situation mirrors a similar event just last month, when fellow firm EY also recalled a report on loyalty rewards programs due to apparently fabricated footnotes and AI-generated inaccuracies. These developments highlight the global imperative for rigorous quality control and human accountability in an era where AI tools are rapidly transforming how professional content is generated and consumed.