KPMG International has officially withdrawn its global report, "Redefining Excellence in the Age of Agentic AI," following widespread challenges from several organizations that disputed the accuracy of case studies featured within. The consulting giant has initiated an internal investigation into the matter, acknowledging potential breaches of its internal AI-use policies.
Report Claims Disputed by Major Organizations
The now-retracted report had asserted that companies and public bodies were already leveraging AI agents to revolutionize operations across diverse sectors like finance, healthcare, and transport. However, these claims were swiftly contested by affected organizations, leading to the report's removal.
Among those who flagged inaccuracies were prominent entities such as UBS, the UK’s National Health Service (NHS), Swiss Federal Railways, and Transport for London. For instance, KPMG had claimed UBS was utilizing AI agents for investment advisory and risk management via a custom Microsoft platform, a claim a UBS spokesperson labeled as “factually incorrect.”
Similarly, Swiss Federal Railways refuted descriptions of its use of AI agents for passenger journey planning, stating the information was “not accurate.” Transport for London found claims about its use of AI agents to predict congestion and coordinate city transport “misleading.” NHS Greater Manchester also challenged KPMG's assertion that it employed AI agents for patient record organization, automated referrals, and predicting hospital readmissions, stating it “doesn’t really align” with reality.
AI Hallucinations and Internal Investigation
The errors were initially brought to light by tech research firm GPTZero and subsequently verified by the Financial Times. It appears that several case studies within the KPMG AI report may have been generated through AI hallucinations, leading to fictitious scenarios being presented as real-world applications.
A KPMG spokesperson confirmed the firm's commitment to the “accuracy and integrity” of its publications and indicated that employees might have violated internal guidelines concerning the responsible use of AI. The firm expects all its personnel to adhere to policies requiring human oversight to validate content and verify independent sources.
This incident closely follows a similar event weeks prior, where EY also withdrew a study that reportedly contained fabricated footnotes and AI-generated errors, a situation also uncovered by GPTZero. These occurrences raise significant questions about the responsible integration of artificial intelligence into research and thought leadership within the consulting industry.