KPMG Pulls AI Report After 89% of Citations Found Inaccurate

Major consulting firm withdraws publication amid accuracy crisis

KPMG has pulled a flagship report on agentic AI from its websites after an independent analysis revealed that 40 of its 45 citations—nearly 89 percent—were inaccurate, misleading, or fabricated. The October 2025 report, titled "Total Experience: Redefining Excellence in the Age of Agentic AI," was intended to showcase cutting-edge AI deployments but instead became an object lesson in the technology's tendency to hallucinate.

Research firm GPTZero conducted what it called a forensic review of the Big Four consulting firm's publication and found that only five citations correctly pointed to their stated sources. The remainder ranged from mangled references to entirely invented titles. GPTZero coined the term "vibe citing" to describe the phenomenon—analogous to "vibe coding"—where generative AI assembles fragments of real sources into references that appear legitimate until verified.

According to GPTZero's analysis, roughly half of the report's factual claims were either false, unsupported, or attributed to incorrect sources. Case studies highlighting supposedly advanced agentic AI deployments at organizations including UBS, Swiss Federal Railways, and Transport for London were particularly problematic, with cited sources either failing to substantiate the claims or containing alterations that undermined their reliability.

Specific errors compound credibility issues

The problems extended beyond footnotes. On page 42, the report claimed Emirates airline deployed a mobile chatbot named Sara capable of conversing with passengers and changing their flights. GPTZero determined that Sara is actually a robot assistant introduced in 2023—not a chatbot—and lacks flight-booking modification capabilities.

The report also contradicted KPMG's own research. It cited 55 percent of CEOs ranking AI as their top investment priority, while KPMG's 2025 CEO Outlook—released the same month—reported that figure as 71 percent.

Why it matters

This incident highlights a critical vulnerability in professional services firms' adoption of generative AI: the technology can produce authoritative-looking content that fails basic fact-checking. For consulting firms that sell trust and expertise, publishing AI-generated fabrications undermines their core value proposition. The episode also demonstrates that even organizations warning clients about AI hallucinations can fall victim to them, suggesting current human oversight processes may be insufficient to catch sophisticated errors at scale.

Industry pattern emerges

KPMG is not alone in this predicament. Last year, Deloitte refunded the Australian government after AI-generated content appeared in a taxpayer-funded report, establishing a troubling pattern within the consulting industry.

A KPMG spokesperson told The Register that the firm "takes the accuracy and integrity of its published content seriously" and is reviewing how the publication was released. The spokesperson emphasized that KPMG expects all personnel to follow guidelines on responsible AI use, including human oversight to validate content and verify independent sources.

The Financial Times first reported that KPMG removed the report from some of its websites while investigating the circumstances surrounding its publication. Details of the citation analysis and specific errors were documented by GPTZero and reported by The Register.