KPMG's AI Report Faces Scrutiny Over Dubious Citations

KPMG's October 2025 report, titled "Total Experience: Redefining Excellence in the Age of Agentic AI," has landed in hot water for allegedly fabricating or misrepresenting a significant portion of its citations. According to the research firm GPTZero, only five out of 45 citations in the report accurately pointed to the intended source. The rest were a mix of misleading, altered, or outright fabricated references.

Unraveling the 'Vibe Citing' Phenomenon

This isn't just another case of minor inaccuracies slipping through the review process. GPTZero has coined the term "vibe citing" to describe this issue, a nod to the way generative AI stitches together fragments of real sources, creating references that appear convincing but don't hold up under scrutiny. It's akin to smoke without fire, and it begs the question: How much can we trust AI-generated content?

The report's claims were so flawed that GPTZero estimates about half of its factual assertions were either false, unsupported, or tied to incorrect sources. Highlights included dubious case studies on agentic AI deployments at UBS, Swiss Federal Railways, and Transport for London. Upon investigation, the sources cited in these examples failed to substantiate the claims or were so heavily altered that their reliability was undermined.

The Consulting Industry's AI Conundrum

This isn't an isolated event in the consulting world. Just last year, Deloitte found itself in a similar predicament, issuing a refund to the Australian government after AI-generated content crept into a publicly funded report.

KPMG's own research appeared contradictory within the report. On one hand, a figure suggested that 55 percent of CEOs ranked AI as their top investment priority. Yet, KPMG's CEO Outlook for the same year put this number at a staggering 71 percent. How can such discrepancies arise? It's a symptom of the larger issue: when AI is left unchecked, it can lead to significant inaccuracies.

Accountability and the Road Ahead

In response to the uproar, KPMG has removed the contentious report from some of its websites while it investigates the publication process. A spokesperson for KPMG stated, "KPMG International takes the accuracy and integrity of its published content seriously." They emphasized the importance of human oversight in validating AI-generated content.

Consulting firms have long cautioned their clients about AI's propensity to "hallucinate" or generate fictional data. Yet, KPMG's recent debacle serves as a stark example of this phenomenon in action. As the industry continues to integrate AI into its operations, one must ask: Are we placing too much trust in AI's capabilities without sufficient oversight? The real world is coming industry, but it's key for human oversight to guide this transition.