KPMG Withdraws AI-Generated Report After Widespread Hallucinations Exposed
Key Takeaways
- ▸KPMG's AI-generated report contained multiple false claims about how organizations were using AI
- ▸GPTZero confirmed the inaccuracies stemmed from AI hallucinations rather than data sourcing errors
- ▸Four major organizations publicly contradicted claims made about them, exposing the report's credibility issues
Summary
Professional services firm KPMG has withdrawn an October 2025 report titled 'Redefining excellence in the age of agentic AI' after research organization GPTZero identified numerous inaccuracies caused by AI hallucinations. Multiple major organizations—including UBS, the UK's National Health Service, Swiss Federal Railways, and Transport for London—reported that claims about their AI usage in the report were either false or misleading. The report appears to have been written with AI assistance but lacked adequate human oversight to validate facts and verify sources. KPMG is conducting an investigation and reiterated its requirement that all employees follow guidelines on responsible AI use, including human verification of content. The incident reflects a growing pattern in the professional services industry of deploying generative AI to produce information-rich content without sufficient fact-checking.
- Similar incidents at EY suggest a broader industry trend of AI misuse without proper human oversight
Editorial Opinion
This is a damning indictment of how enterprises are deploying generative AI. Professional services firms like KPMG have reputational stakes that depend on accuracy, yet they appear to be cutting corners by letting AI generate claims without rigorous fact-checking. Using LLMs to author reports about technology practices without human verification isn't just negligent—it undermines trust in both the firms and the technology itself.
