KPMG pulls report on AI usage due to apparent hallucinations
KPMG removed a report after multiple organizations said its claims about their AI use were untrue or misleading.
Intelligence analysis by GPT-5.4 Mini
KPMG withdrew a report on agentic AI after outside groups and the Financial Times raised concerns that parts of it were inaccurate. The firm says it is investigating and reaffirmed that AI-generated content needs human review and source verification.
KPMG published a report that talked about how other groups use AI, but some of those groups said the report got facts wrong. It is like a class paper that looks neat but has made-up sources, so the teacher has to take another look.
Analysis
KPMG pulled a report titled “Redefining excellence in the age of agentic AI” after several organizations said the report misrepresented their AI usage. According to TechCrunch, research group GPTZero found multiple inaccuracies and told the Financial Times that the errors appeared to come from AI hallucinations.
The reported mistakes mattered because they involved claims about real companies and institutions. UBS, the UK’s National Health Service, Swiss Federal Railways, and Transport for London all told the FT that the report’s statements about their AI usage were untrue or misleading. That made the problem bigger than a small citation mistake: the report appears to have overstated or fabricated evidence about how widely AI was being adopted.
KPMG said it removed the report from its websites while it investigates. A spokesperson said the firm expects employees to follow responsible AI guidelines, including human oversight to validate content and verify independent sources. That framing suggests KPMG is treating the issue as both a process failure and a reputation problem.
The episode lands soon after another AI-related publishing mistake at EY, which withdrew a report on loyalty rewards programs that reportedly contained fake footnotes and AI hallucinations. Together, the cases show a recurring pattern: AI can speed up drafting, but it can also produce confident-sounding errors that survive into public-facing reports unless people check carefully.
Key points
- KPMG removed a report on agentic AI after organizations said its claims about their AI use were false or misleading.
- GPTZero identified inaccuracies and said they appeared to come from AI hallucinations.
- UBS, the UK NHS, Swiss Federal Railways, and Transport for London disputed the report’s claims.
- KPMG said it expects human oversight to validate content and verify independent sources.
- The case follows a similar EY report withdrawal over fake footnotes and hallucinations.
If KPMG’s review leads to stronger checks, the episode could push firms to use AI more carefully in research and reporting. That could make future reports more trustworthy by forcing human verification before publication.
If companies keep relying on AI drafts without enough review, more public reports could contain false claims and damage trust. The incident could also make readers more skeptical of AI-related research, even when it is accurate.
