KPMG pulls AI report after review finds hallucinated claims

KPMG has pulled an artificial intelligence report after an outside review found apparent hallucinations, false claims and misattributed citations in a document about AI adoption. The contradiction is striking: a major consulting firm had to remove a report on AI because the sourcing did not survive scrutiny. The episode now raises a sharper question for corporate research and policy advice, what verification standard should govern material that can influence boardrooms and regulators?

The report, titled Total Experience: Redefining Excellence in the Age of Agentic AI, was published in October 2025. GPTZero said it reviewed 45 citations and found only five that correctly pointed to the cited source. The group said roughly half of the report’s factual claims were false, unsupported or attributed to the wrong source, while another 28 citations relied on paraphrased titles that were misleading or too vague to verify.

Several of the disputed claims centered on high-profile organizations including UBS, Swiss Federal Railways, Transport for London and Emirates. GPTZero flagged one passage that described an Emirates mobile chatbot named Sara as able to converse with passengers and change flights. GPTZero said Sara is actually a robot assistant introduced in 2023 and cannot alter bookings. It also said the report conflicted with KPMG’s own 2025 CEO Outlook, which put the share of CEOs ranking AI as their top investment priority at 71 percent, not 55 percent as the report claimed.

KPMG removed the report from some of its websites while it reviews the circumstances surrounding its publication. A KPMG spokesperson said, "KPMG International takes the accuracy and integrity of its published content seriously." GPTZero described the citation problem as "vibe citing," a phrase for references that sound plausible but prove inaccurate, fabricated or impossible to verify.

The fallout extends beyond one report and one firm. The episode feeds a broader concern in consulting and advisory circles: AI-generated or AI-assisted content can move quickly from draft to polished thought leadership without adequate source checking. When a firm of KPMG’s stature publishes research on AI adoption, the stakes are larger than reputation alone. Clients, investors and policymakers rely on that material to judge how fast enterprises are adopting the technology, where the risks lie and what governance standards are needed next.