76Trust
Verified
🔍 Web Verified
Kevin BeaumontonMastodon1d ago
KPMG issued a report citing all the transformational ways GenAI has transformed industry, it’s been widely cited.
One minor problem: it turns they used AI to write the report, and it made up all of the evidence.
KPMG have now withdrawn the report in full.
https://www.ft.com/content/b3828e92-4961-4b39-84f0-c42f33be3c3f
Trust Metrics
80
72
70
75
Accuracy80%
Framing72%
Context70%
Tone75%
Analysis Summary
KPMG published an October 2025 report on AI benefits that was heavily promoted, but a subsequent investigation by GPTZero (reported by CityAM) found the report was full of AI hallucinations — of 45 citations checked, only 5 pointed to real, uncorrupted sources. KPMG's use of AI to research a report about AI creating fake evidence is a real problem because it exposes how difficult it is for professional firms to catch hallucinations at scale, and it undermines trust in AI adoption reports that enterprises rely on for decision-making. The broader issue is that KPMG and other Big Four firms are actively using AI in their own consulting work while simultaneously selling AI adoption expertise to clients — creating a conflict of interest in their ability to credibly assess AI risks.
Claims Analysis (3)
“KPMG issued a report citing transformational ways GenAI has transformed industry, widely cited”
Multiple sources confirm KPMG published an October 2025 report on agentic AI titled 'Total Experience: Redefining Excellence in the Age of Agentic AI' that received significant attention.
“KPMG used AI to write the report and it made up all of the evidence”
Multiple sources confirm KPMG used AI in report preparation and that it contained fabricated/hallucinated citations. CityAM reports GPTZero found only 5 of 45 citations were accurate. 'All of the evidence' is slightly overstated — some citations were valid, but widespread hallucination is confirmed.
“KPMG has now withdrawn the report in full”
Search results confirm the report faced scrutiny and KPMG acknowledged errors, but explicit confirmation of 'full withdrawal' is not clearly stated in available sources. CityAM and other outlets confirm the citation problems and investigation, but formal withdrawal language is not captured in the search results provided.
Was this analysis helpful?
Try ClearFeed free →