KPMG published a report on AI agents with hallucinations: fabricated quotes and non-existent cases

KPMG — one of the world's largest consulting companies — released a report on artificial intelligence agents in October 2025. The problem is that the document itself appears to have been written with the help of AI: it contains fabricated quotes and references to nonexistent case studies.

What Was Found in the Document

Researchers who examined KPMG's publication identified several types of errors characteristic of AI hallucinations. The text contained quotes attributed to real experts that those experts never actually made. Some references pointed to nonexistent studies — materials that do not appear in any archives. Here is a typical set of problems that the reviewers uncovered:

• Quotes from real people with fabricated content
• References to nonexistent case studies
• Statistics without verifiable sources
• Case studies of real companies with fictional details
• Mentions of research that cannot be found in public databases

Such errors are well-known to AI specialists: these are classic hallucinations of language models — confident generation of plausible but fabricated content.

Why This Is Painful for the "Big Four"

KPMG is part of the "Big Four" of global audit and consulting firms along with Deloitte, PwC, and Ernst & Young. Their key competitive advantage is accuracy, reliability, and authority. Corporate clients pay tens of thousands of dollars for research precisely because they trust its quality. The situation becomes especially ironic when you consider the context: KPMG actively sells consulting services to clients on AI implementation and warns about the risks of thoughtlessly applied AI tools. The scandal showed that the company itself fell into the very trap it warns others about.

How Does This Even Happen

The exact circumstances are not disclosed, but the scenario is quite typical. An analyst uses a language model to speed up the writing of research — this is normal practice in 2025. The problem arises at the verification stage: if the author does not manually verify each quote and each reference, hallucinations go unnoticed. Given high demand for AI expertise, deadlines are often tight, and the temptation to delegate routine work to the model is great. But for materials where accuracy determines reputation, such savings on verification prove costly.

"When a company that consults on AI risks releases a report with AI

hallucinations — this is the best illustration of the risk itself."

What Does This Mean

The KPMG incident becomes a warning for the entire industry: using AI to create professional reports requires mandatory manual verification of quotes and sources. A brand's reputation does not replace fact-checking — and the higher the trust in the author, the more destructive the consequences when that trust is not justified.