KPMG pulls report on AI usage due to apparent hallucinations
Original reporting by TechCrunch
Professional services giant KPMG has been compelled to withdraw a prominent report, "Redefining excellence in the age of agentic AI," after numerous organizations refuted its claims about their AI usage. The October 2025 publication, intended to showcase cutting-edge AI adoption, instead became a stark illustration of AI's pitfalls, with research group GPTZero identifying widespread inaccuracies linked to "AI hallucinations." This implies a significant irony: KPMG appears to have used AI to help write a report *about* AI, only for that AI to generate fabricated information.
UBS, the UK's National Health Service, Swiss Federal Railways, and Transport for London were among those who informed the Financial Times that the report's assertions regarding their AI implementations were either untrue or misleading.
Responding to missteps
A KPMG spokesperson confirmed the report's removal pending an internal investigation, emphasizing the firm’s commitment to "human oversight to validate content and verify independent sources" in its AI guidelines. This incident echoes a similar withdrawal last month by fellow Big Four firm EY, which pulled a report riddled with fake footnotes and apparent AI-generated falsehoods. These events underscore the urgent need for robust human vetting as professional services increasingly integrate generative AI into their research and reporting workflows.
KPMG's swift withdrawal of its "Redefining excellence" report, following similar action by EY last month, serves as a stark reminder of the volatile intersection between ambitious AI integration and fundamental content integrity. The irony that a report ostensibly analyzing agentic AI appears to have succumbed to the very AI hallucinations it aimed to explore is particularly biting. This incident not only calls into question the firm's internal vetting processes but also highlights the critical gap that can emerge when the pursuit of efficiency, facilitated by AI, overshadows the imperative of accuracy and robust human oversight, a standard KPMG itself explicitly champions.
The Trust Imperative
The fallout from these events extends far beyond individual firms, signaling a burgeoning crisis of credibility for any organization rushing to adopt generative AI for content creation without rigorous safeguards. Professional services, whose very value proposition rests squarely on expertise, trustworthiness, and reliable advice, face magnified reputational risks that can take years to rebuild. This episode underscores that while AI offers powerful tools for data synthesis and content generation, it simultaneously demands an even greater commitment to human verification, critical thinking, and comprehensive source validation. As AI capabilities rapidly advance, the industry must not only refine its technological infrastructure but, more importantly, fortify its ethical frameworks and quality control mechanisms to prevent such fundamental errors. The future impact will likely see increased scrutiny from clients and regulators, heightened internal training on responsible AI use across all sectors, and potentially more stringent industry-wide standards, all aimed at restoring and maintaining the bedrock of trust in an increasingly AI-infused world.
Frequently asked questions
- Why did KPMG withdraw its recent report on AI usage in professional services?
- Professional services firm KPMG retracted its "Redefining excellence in the age of agentic AI" report after multiple organizations, including UBS and the NHS, denied its claims about their AI implementations. The report contained widespread inaccuracies, attributed to "AI hallucinations," suggesting that AI tools used in its creation generated fabricated information. This incident underscores the critical need for rigorous human oversight and verification in AI-assisted content generation.
- What are AI hallucinations and how do they impact professional reports and research?
- AI hallucinations occur when generative AI models produce false, misleading, or nonsensical information, presenting it as factual. In professional reports, this can lead to fabricated data, incorrect assertions, or non-existent sources, as seen in recent incidents involving major firms. Such errors undermine credibility, necessitate report retractions, and highlight the essential role of human fact-checking and critical analysis to ensure accuracy.
- What are the key implications for professional services using AI in content creation?
- The use of AI in content creation by professional services firms carries significant reputational risks if not managed with robust safeguards. Incidents of AI-generated inaccuracies underscore the imperative for enhanced human oversight, stringent content validation, and comprehensive source verification. Maintaining client trust requires fortifying ethical frameworks and quality control mechanisms, potentially leading to increased scrutiny from clients and regulators, and more rigorous industry standards.