A Consulting Firm Published a Report Full of Made-Up Examples. Here's Why It Matters.

KPMG, one of the world's largest consulting firms, has pulled a report about artificial intelligence from its websites after discovering it contained false information. The report, called Redefining Excellence, included made-up case studies, fake citations, and claims about companies using AI that those companies never actually made.

The errors were spotted by a research group called GPTZero and later confirmed by the Financial Times. The fake examples included stories about how banks like UBS and various hospitals and transit systems were supposedly using AI. When contacted, these organizations said the descriptions of their AI use were completely false. UBS specifically asked KPMG to remove any mention of the firm from the report.

KPMG is now investigating what went wrong. According to The Register, the company has not announced a timeline for finishing that investigation or how it will move forward.

What Happened Here

To understand why this matters, it helps to know what AI hallucination is. When you use an AI chatbot or language model, it generates text word by word, always trying to predict what should come next. Sometimes it produces sentences that sound convincing but contain facts that are simply wrong — it invents citations, statistics, or examples that never existed. This is called hallucination. It happens because the AI is good at mimicking the style of language but has no real understanding of whether what it is saying is true.

KPMG's report was supposed to be about the risks of AI hallucination. The irony is obvious: the report itself became a live example of that exact problem.

Why This Is Actually a Bigger Problem Than It Sounds

Here is the key point worth stopping on: AI hallucination is not the real failure here. The failure was that humans did not catch it before the report went public.

When companies use AI to help write documents, the output still needs human review. Someone should read it, check the facts, confirm that the citations are real, and verify that companies mentioned in case studies actually did what the report claims. That did not happen at KPMG, or it happened poorly enough to miss multiple fabricated examples tied to real, named organizations.

For KPMG in particular, this is a serious problem. The firm's entire business is built on trust. Clients pay KPMG specifically because they expect rigorous, accurate information and careful analysis. A published report full of invented examples named after real companies — companies that then had to contact KPMG to complain — directly undermines that promise.

The downstream effect is worth watching. Many consulting firms and research companies have started using AI tools to speed up report writing. AI can help with drafting, summarizing data, and other tasks that save time. But the old safeguards — having editors review everything, checking sources carefully, making sure any real company names are accurate — have not always kept up with the rush to use these tools faster. This incident will likely force firms to slow down and build better review processes.

What To Do About This If You Read Reports Like This

If you see a research report or analysis that uses AI, and it includes specific examples, statistics, or case studies with real company names attached, do yourself a favor: check those claims independently before you trust them. AI-generated writing can sound very natural and persuasive, even when it is wrong. That smooth, confident tone is exactly what makes false information harder to spot when you are skimming a document. The extra step of verification is not optional.