KPMG’s AI Debacle: When Aspirations Collide with Illusions
KPMG’s report from October 2025 on the marvels of agentic AI has found itself in deep trouble, unveiling one of AI’s less appealing traits: fabrication. Researchers at GPTZero believe a thorough examination of the Big Four firm’s publication, “Total Experience: Redefining Excellence in the Age of Agentic AI,” revealed that a mere five out of 45 citations were actually accurate. The remaining citations? A blend of distorted details, deceptive information, or too vague to clarify.
The Chaotic Realm of “Vibe Citing”
It appears that the consulting sector has experienced this before. Last year, Deloitte had to refund the Aussies after a government-funded report included some AI-generated fluff. GPTZero has dubbed this fiasco “vibe citing” – where AI compiles fragments from actual sources, invents titles, or produces references that seem impressive until scrutinized.
Half-Formed Assertions and AI Myths
GPTZero asserts that about 50% of the report’s factual statements were either false, unsubstantiated, or outright misattributed. Some of the allegedly top-tier agentic AI case studies resembled more of a creative writing endeavor than truth. Highlighted examples featured so-called AI innovations at UBS, Swiss Federal Railways, and Transport for London. According to GPTZero, the sources did not support the claims and included alterations that rendered them questionable.
Case Study Shenanigans
“These inaccuracies aren’t merely hiding in the report’s footnotes,” GPTZero exclaimed. “On page 42, the authors claim Emirates airline has a mobile chatbot named Sara (nonsense) that interacts with passengers directly (some truth) and modifies their flights (false). Sara is actually a robotic companion introduced in 2023 (not a chatbot) and cannot alter flight reservations.”
Internal Discrepancies Abound
And it’s not just external references that are getting the KPMG twist. GPTZero pointed out a contradiction with KPMG’s own findings, stating that 55% of CEOs ranked AI as their primary investment priority. However, KPMG’s 2025 CEO Outlook from the same time indicated it was 71%. Quite the puzzler.
KPMG’s Remediation Effort
KPMG has retracted the report from several platforms while investigating how this disaster made it through, as reported by the Financial Times. A KPMG spokesperson informed GadgetLad: “KPMG International places great importance on the accuracy and integrity of its published material. The report has been removed, and we’re exploring how it was released. We expect our team to adhere to guidelines on responsible AI usage, including having a human review to verify content and consult independent sources.”
Learning from AI’s Exaggerated Tales
Consulting firms have long cautioned clients about AI fabricating information. It seems KPMG may have just given everyone an up-close view.
Summary
When AI Whimsies Get Bold – Who could have predicted KPMG’s report would turn into a prime illustration of AI’s imaginative storytelling? It’s a classic case of ‘practice what you preach’, right?
