Imagine a world where a top consulting firm’s big AI report turns out to be a house of cards. That’s exactly what happened. One of the biggest names in consulting released a detailed report on AI benefits. But the truth? The report was riddled with fake data and hallucinated facts created by AI itself.

The Hallucination Storm Hits Big Consulting

KPMG, a giant in professional services, aimed to showcase how AI is revolutionizing customer experience. Their report, “Total Experience: Redefining Excellence in the Age of Agentic AI,” promised cutting-edge insights. Instead, it became a cautionary tale about AI’s pitfalls. An independent investigation revealed that most of the report’s references were fake or wildly inaccurate. Out of 45 citations, only five pointed to real sources. The rest? Fabricated or distorted beyond recognition.

How does this happen? The culprit is a phenomenon called “vibe citing.” It’s when generative AI models invent believable but non-existent references to back up claims. This isn’t just a typo or a minor mistake. It means large chunks of the report’s foundation were built on thin air.

Worse still, about half of the factual claims in the report were false or misattributed. For example, the report claimed a Swiss bank used agentic AI to manage investments and compliance. The bank denied this completely. Another claim about a mobile chatbot that could reroute flights turned out to be false. These weren’t careless errors—they were AI-made fabrications that slipped through unchecked.

Why This Matters: The Domino Effect of False AI Reports

Consulting firms like KPMG hold immense influence. Their research shapes business decisions worldwide. When they drop the ball, misinformation spreads fast. Industry publications and even AI tools like ChatGPT have started citing these faulty reports. That creates a dangerous feedback loop. AI learns from bad data, then generates more misinformation.

This isn’t the first time a Big Four firm stumbled over AI hallucinations. Deloitte had to refund a government contract after similar errors. EY pulled a cybersecurity report for fake citations. Even elite law firms have faced backlash for AI-generated false legal documents. The pattern is clear: the rush to use AI without proper checks is causing costly mistakes.

• Fake citations can poison trusted knowledge pools.
• Companies risk making decisions based on false data.
• Unchecked AI can mislead not just reports but entire industries.
• Human oversight remains crucial to verify AI-generated content.

Lessons Learned and What Comes Next

KPMG has since removed the report and is investigating how this happened. The firm admitted no one double-checked those AI-generated citations or claims before publishing. This highlights a glaring gap in quality control. AI tools can speed up research, but humans must remain gatekeepers.

What does this mean for businesses and AI users? It’s a wake-up call. AI can generate impressive content but can also invent convincing lies. Companies must blend AI’s power with strict human review. Transparency about AI’s role in content creation will become central. Some regulators are already pushing for clearer disclosures and audits of AI-generated research.

The broader consulting industry faces a turning point. If trusted advisors can’t catch AI errors, clients may lose faith in their guidance. AI’s promise is huge, but so are its risks. Responsible use, thorough validation, and ethical AI practices are no longer optional—they’re mandatory.

In the end, this scandal serves as a vivid reminder: AI doesn’t just amplify knowledge. It can also amplify errors. The future belongs to those who harness AI wisely, with eyes wide open. Are you ready to separate real insights from AI’s illusions?