A report about artificial intelligence is now facing scrutiny for allegedly containing AI-generated inaccuracies.
And that’s exactly why this story is attracting so much attention across the business and technology world.
Professional services giant KPMG has withdrawn a report focused on agentic AI after multiple organizations challenged claims made about their use of artificial intelligence, raising uncomfortable questions about how companies are using AI to write about AI itself.
What Happened?
The report, titled “Redefining Excellence in the Age of Agentic AI,” was originally published in October 2025.
The controversy emerged after research group GPTZero identified what it described as numerous inaccuracies in the document. According to the Financial Times, GPTZero concluded that the questionable content appeared to be the result of AI hallucinations — instances where AI systems generate information that sounds convincing but is inaccurate or entirely fabricated.
The allegations quickly drew responses from several high-profile organizations named in the report.
Among them were:
- UBS
- The UK’s National Health Service (NHS)
- Swiss Federal Railways
- Transport for London
According to the Financial Times, those organizations said claims regarding their AI usage were either untrue or misleading.
That immediately shifted the conversation from a routine industry report to a much larger debate about trust, verification, and accountability in the age of generative AI.
But that’s only part of the story.
Must Read: 2026’s Controversial AI Crackdown: Did Amazon’s Warning Trigger Anthropic’s Biggest Crisis Yet?
Why This Matters
For months, businesses around the world have been promoting AI-powered productivity gains.
At the same time, executives have repeatedly stressed the importance of human oversight.
The KPMG situation puts those promises under a spotlight.
A spokesperson for KPMG said the firm removed the report from its websites while conducting an internal investigation.
The spokesperson also emphasized that employees are expected to follow the company’s responsible AI guidelines, including human review of content and verification of independent sources.
That statement highlights a growing tension inside corporate AI adoption:
AI can accelerate research and content creation.
But when verification breaks down, the risks can become public very quickly.
Key Takeaway
| Issue | Potential Risk |
|---|---|
| AI-generated inaccuracies | Misleading readers |
| Weak verification processes | Reputational damage |
| Incorrect corporate claims | Loss of trust |
| Hallucinated content | Questions about oversight |
And this is where reactions started spreading across the industry.
A Growing Pattern?
The KPMG incident is not occurring in isolation.
Last month, EY withdrew a separate report on loyalty rewards programs after concerns emerged that it appeared to contain fake footnotes and AI hallucinations.
While the circumstances differ, both cases have fueled broader concerns about how organizations are integrating generative AI into research, analysis, and publication workflows.
The irony has not gone unnoticed.
Two major professional services firms now find themselves defending reports that allegedly suffered from the very AI reliability issues many businesses are still trying to understand.
That raises a difficult question:
If large organizations with established review processes can encounter these problems, how widespread could similar issues be elsewhere?
The Hidden Problem Behind AI Hallucinations
AI hallucinations remain one of the most persistent challenges facing generative AI systems.
The outputs can appear authoritative, polished, and highly convincing.
That’s precisely what makes them dangerous.
Readers often assume professionally formatted reports have been thoroughly verified.
But when inaccurate information slips through, the appearance of credibility can amplify the impact of mistakes.
In highly regulated industries, that concern becomes even more significant.
A report isn’t just content.
It’s often used to influence decisions, investments, policies, and public understanding.
Contrarian View: Is AI Really the Main Problem?
Not everyone will view this controversy as an AI failure.
A different perspective is that the underlying issue may be human process rather than technology itself.
Supporters of this argument note that AI systems are tools, not autonomous decision-makers. If inaccurate information made it into a published report, critics of the “AI is to blame” narrative may argue that review and validation procedures deserve equal scrutiny.
In other words, the debate isn’t only about what AI generated.
It’s also about what humans approved.
That distinction could become increasingly important as AI-assisted research becomes more common across consulting, finance, healthcare, transportation, and government sectors.
What Happens Next?
KPMG says it is conducting its own investigation.
The findings could provide important insight into how the inaccuracies entered the report and whether existing safeguards worked as intended.
More broadly, the episode may become another case study in a debate that is far from settled.
Companies continue racing to adopt AI tools.
Yet every high-profile hallucination controversy reminds organizations of the same reality:
Speed is valuable, but credibility is harder to rebuild once it’s questioned.
And as AI-generated content becomes increasingly common in boardrooms, consulting firms, research reports, and public communications, one question is likely to linger:
How much verification is enough when even trusted institutions can get caught by information that never should have been there?
Editorial Disclaimer: This article is based entirely on publicly available information reported by TechCrunch, the Financial Times, statements attributed to involved organizations, and other cited sources. No facts, quotes, outcomes, or timelines have been fabricated. Analysis and interpretation may evolve as additional information becomes available.