KPMG report contains fabricated citations and errors
CityAM reports that GPTZero's investigation into KPMG's October 2025 report "Total Experience: Redefining Excellence in the Age of Agentic AI" found widespread citation problems. Per CityAM, of 45 citations in the report, only five accurately point to real, uncorrupted sources; GPTZero flagged 40 citation titles as fabricated. CityAM reports roughly half the factual claims appear false or misattributed, including a misused 2019 East Japan Railway press release and a numerical contradiction: the report cites KPMG research claiming 55% of CEOs rank AI as their top investment priority while KPMG's own 2025 CEO Outlook published the same month shows 71%. KPMG told CityAM it "takes the accuracy and integrity of its published content seriously" and removed the report from its homepage while investigating. GPTZero researcher Paul Esau is quoted: "We suspect no human at KPMG double-checked the citations, the claims, or the sources before Total Experience was published" (CityAM).
What happened
CityAM reports that GPTZero probed KPMG's October 2025 report "Total Experience: Redefining Excellence in the Age of Agentic AI." Per CityAM, the investigation found that of 45 citations, only five accurately point to real, uncorrupted sources; GPTZero flagged 40 citation titles as fabricated. CityAM reports roughly half the factual claims supported by those citations appear false or misattributed. Examples cited include a 2019 East Japan Railway press release erroneously used as evidence of early agentic AI adoption, and repeated errors where KPMG's language model attributed articles to their subjects rather than their actual authors (CityAM). CityAM also notes an internal contradiction: the report cites "KPMG research" claiming 55% of CEOs rank AI as their top investment priority, while the KPMG 2025 CEO Outlook published the same month records 71% (CityAM). GPTZero researcher Paul Esau is quoted: "We suspect no human at KPMG double-checked the citations, the claims, or the sources before Total Experience was published" (CityAM). KPMG told CityAM it "takes the accuracy and integrity of its published content seriously" and has removed the report from its homepage while investigating.
Wider implications
CityAM notes that some of the report's flawed material has been republished by industry outlets and is now surfacing when users query ChatGPT and Gemini, illustrating how a single fabricated-citation source can propagate through downstream media and language models. GPTZero has run similar investigations into other consulting firms' AI-assisted reports. For practitioners, this episode reinforces that AI-assisted research and drafting pipelines require explicit human verification of citations before publication, particularly for externally distributed analysis where errors can compound through reuse.
Scoring Rationale
The incident is notable for practitioners: a Big Four firm published an AI-assisted report with an 89% citation-fabrication rate, confirmed by GPTZero and corroborated by CityAM. The cascade effect - flawed material spreading to industry media and LLM outputs - adds material practitioner implications. Single investigative source with limited global AI-ecosystem impact places this at the high end of Solid rather than Notable.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
