KPMG has retracted its global report on “Agentic AI” after receiving serious complaints from several prominent companies. These companies claimed that the report contained completely fabricated achievements attributed to them. The consulting giant had released the report, titled “Redefining Excellence in the Age of Agentic AI,” which was later found to contain misleading claims and false case studies, as revealed by The Financial Times. The inaccuracies appeared to stem from what some are calling “AI hallucinations,” where artificial intelligence produces incorrect information that human reviewers failed to detect.
The misinformation was initially highlighted by the tech research group GPTZero and subsequently confirmed by The Financial Times. Many organizations, including UBS, the UK’s National Health Service (NHS), and various public transit agencies, pressured KPMG to remove the report from its website after the inaccuracies were identified.
Claims and Responses
KPMG’s report suggested that top global organizations had successfully integrated advanced AI agents for complex, automated tasks. However, the companies involved were quick to clarify the claims:
-
UBS Bank: The report argued that UBS integrated AI across its advisory and risk management systems. A UBS spokesperson called the claims “factually incorrect” and insisted on their removal.
-
Swiss Federal Railways: KPMG suggested that the railways used AI agents to help users plan and book journeys. A spokesperson stated these claims were “not accurate.”
-
Transport for London (TfL): The report claimed that London’s transport system used AI agents to predict congestion. TfL called this claim “misleading.”
-
NHS Greater Manchester: The report stated that the NHS utilized AI agents for patient data management. A representative responded that this did not align with actual practices.
Erosion of Trust
This incident is not isolated; just last month, another major firm, EY, retracted a study after similar issues of AI-generated errors were uncovered. The Big Four accounting and consulting firms are typically regarded as reliable, making these oversights particularly damaging. Edward Tian, CEO of GPTZero, highlighted that such errors can undermine trust in information sources, noting that KPMG’s incorrect findings had already been cited by various tech publications and a major European newspaper before the retraction.
KPMG’s Next Steps
KPMG has stated that it takes the accuracy of its content seriously and has pulled the report while initiating a full internal investigation. The firm acknowledged that some employees may have breached internal guidelines regarding the use of AI. A spokesperson emphasized the importance of human oversight to validate AI-generated content and verify sources independently.
