KPMG has retracted a report titled “Redefining excellence in the age of agentic AI” after multiple organizations, including UBS and the UK’s National Health Service, confirmed that claims regarding their AI adoption were inaccurate. Research group GPTZero identified these errors as AI hallucinations, suggesting the professional services firm relied on generative AI to produce content without sufficient human oversight.
AI hallucinations occur when large language models generate false or misleading information while presenting it as fact. This phenomenon poses significant risks for corporate reporting, where accuracy is essential for brand reputation.
Why are professional services firms publishing inaccurate AI reports?
The recent retraction at KPMG follows a similar incident last month involving EY, which withdrew a report on loyalty rewards programs after it was found to contain fake footnotes and AI-generated errors. According to reports from the Financial Times, these incidents highlight a systemic struggle: firms are attempting to demonstrate expertise in AI by using the technology itself, often bypassing necessary human verification. When organizations prioritize speed of publication over rigorous fact-checking, they risk disseminating “hallucinated” data that can damage client relationships and public trust.
What are the consequences of using AI for corporate research?
When firms publish unverified AI-generated claims, they invite immediate public scrutiny from the entities they cite. UBS, the UK’s National Health Service, Swiss Federal Railways, and Transport for London officially disputed the claims made in the KPMG report. A KPMG spokesperson stated the firm is currently conducting an internal investigation and has reinforced guidelines requiring human oversight to validate content and verify independent sources. For these firms, the reputational fallout includes the loss of credibility with industry peers and a heightened need to prove their internal governance processes work.
How can companies avoid AI-driven reporting errors?
Industry standards for AI-assisted writing are shifting toward a “human-in-the-loop” model. Experts suggest that firms must treat AI as a drafting tool rather than an authoritative source of truth. To mitigate risks, organizations should implement the following:
- Mandatory Fact-Checking: Every AI-generated claim must be cross-referenced against primary, non-AI sources.
- Disclosure Policies: Transparency regarding the extent of AI involvement in research reports.
- Human Oversight: Establishing clear accountability for final document accuracy at the senior management level.
If you use AI to summarize industry trends, always treat the output as a draft. Use search engines to verify specific company claims or data points before finalizing your report.
Frequently Asked Questions
What is an AI hallucination?
An AI hallucination occurs when a generative AI model produces incorrect or fabricated information while maintaining a confident, professional tone.
Why did KPMG pull its report?
KPMG removed the document because several organizations, including the UK’s National Health Service and UBS, stated that the report’s claims regarding their specific AI usage were untrue.
Are other firms facing similar issues?
Yes. EY withdrew a report last month after it was found to contain fake footnotes and AI-generated inaccuracies.
How can readers verify AI-generated reports?
Always look for citations that link to primary sources. If a report makes specific claims about a company, check that company’s official press releases or verified news coverage to confirm the information.
Have you encountered errors in AI-generated reports in your industry? Share your experiences in the comments below or subscribe to our newsletter for more updates on the responsible use of AI in corporate communications.
