Particle.news

KPMG Withdraws Agentic AI Report After Multiple Case Studies Are Shown To Be False

The firm has removed the paper and opened an internal probe as independent checks found AI-generated fabrications and named clients demanded corrections.

Overview

  • KPMG pulled its global report, Redefining Excellence in the Age of Agentic AI, after organisations named in the paper disputed specific operational claims and asked for them to be removed.
  • Errors in several case studies were first flagged by AI-detection firm GPTZero and then corroborated by reporting from the Financial Times, which found some examples appeared to be AI hallucinations.
  • Clients named in the report, including UBS, NHS Greater Manchester, Swiss Federal Railways and Transport for London, publicly rejected KPMG’s descriptions and called the claims inaccurate or misleading.
  • KPMG said it has launched an internal investigation and warned staff may have broken the firm’s AI-use rules while stressing the need for human oversight and source verification in published research.
  • The episode follows a recent EY withdrawal for similar AI-related errors and raises wider risks for consultancies, with likely consequences for how firms verify research, manage client relations and formalise AI governance.