KPMG Report Under Fire for Fabricating AI Case Studies

A major scandal has erupted in the consulting world after KPMG released a report containing entirely fabricated AI case studies. The report, titled "Redefining excellence in the age of agentic AI," falsely attributed artificial intelligence implementations to several high-profile global organizations.

Fabricated Success Stories and Disputed Claims

The controversy centers on KPMG's attempt to demonstrate the value of "agentic AI" through real-world examples. However, investigations by GPTZero and the Financial Times revealed that the claims regarding AI usage at UBS, the UK's National Health Service (NHS), Swiss Federal Railways, and Transport for London were completely untrue.

Upon being confronted with the findings, all the named organizations disputed the claims, stating that the purported AI integrations described in the report never took place. This failure strikes at the heart of professional consulting, where empirical evidence and verified case studies are the primary currency used to build client trust and justify massive digital transformation investments.

The Danger of "Secondary Hallucinations"

Beyond the immediate embarrassment for KPMG, the incident highlights a growing systemic risk in the AI ecosystem. GPTZero CEO Edward Tian warns that when reputable consulting firms publish flawed reports, they trigger "secondary hallucinations." Because these documents are considered highly credible, they are frequently ingested, recycled, and redistributed by both human professionals and LLM training sets.

When an AI model learns from a hallucinated report, it propagates that error into future outputs, creating a feedback loop of misinformation that becomes increasingly difficult to untangle. This creates a "hallucination debt" that threatens the reliability of the entire knowledge economy.

"Vibe Citing" and the Failure of AI-Assisted Research

The root cause of the debacle appears to be a lack of rigorous human oversight in the research process, specifically regarding the careless use of AI search tools. GPTZero identified a pattern of "vibe citing"—a phenomenon where citations are merely loose paraphrases of real sources, often lacking correct authors or working URLs. In several instances, the research pointed to original sources that did not exist at all.

This pattern of sloppy sourcing mirrors the issues currently plaguing Google's AI Overviews, which recently faced legal scrutiny in a German court. For a firm like KPMG, which is actively selling AI adoption strategies to global enterprises, the inability to manage the nuances of AI-generated research suggests a significant internal gap in technical literacy and quality control.

Key Takeaways