KPMG removed a report containing examples of the use of artificial intelligence after the accuracy of some of the information it contained was questioned. It was claimed that the report, published in October 2025 and titled “Redefining Excellence in the Age of Agentic AI”, contained inaccurate or misleading statements about the artificial intelligence applications of different institutions. After evaluations on the issue, the company announced that it removed the report from its websites and launched a comprehensive investigation.
In the incident, which was first brought to the agenda by the Financial Times, GPTZero, which conducts studies in the field of artificial intelligence content detection, shared remarkable findings. The company stated that some examples in the report were not supported by verifiable sources and that there were various inconsistencies in the content. According to GPTZero, a significant portion of these inaccuracies may be caused by errors called “hallucinations” that generative artificial intelligence systems can occasionally produce. This situation brought about discussions that artificial intelligence tools may have been used in the creation of a report about artificial intelligence.
Some major organizations mentioned in the report also objected to the claims in the content. Agencies including UBS, the UK NHS, the Swiss Federal Railways and Transport for London confirmed to the Financial Times that information about their use of AI in the report was false or misleading. These statements raised questions about the reliability of the report. The accuracy of case analyses, especially those included in corporate reports, is of critical importance for both industry professionals and decision makers.
Artificial intelligence-related errors in corporate reports are on the agenda again
In the statement made by KPMG, it was emphasized that company employees are expected to comply with the rules regarding the use of artificial intelligence. The company spokesperson stated that all content must be subject to human review and verified by independent sources and that the review process is ongoing. Accordingly, the company is trying to determine at what stages an error occurred during the preparation of the report.
This incident follows recent similar examples in the consulting and professional services sector. In May, EY was also forced to retract a report on customer loyalty programs. It was claimed that the report in question contained fake footnotes and false information thought to be caused by artificial intelligence. Thus, two of the world’s leading consultancy companies soon came to the fore with the risks of artificial intelligence-supported content production.
Generative AI tools are increasingly being used in corporate research, reporting and content creation. However, it has been known for a long time that the information produced by these systems may not always be accurate. Errors that occur especially in areas such as citing sources, verifying data, and creating examples of real institutions show that human control remains important. The developments surrounding the report withdrawn by KPMG also bring to the agenda the verification processes that should be applied in the use of these technologies, as well as the role of artificial intelligence in corporate workflows. For this reason, the view that companies should operate multi-layered control mechanisms before publishing artificial intelligence-supported content is gaining ground.