The consulting industry’s AI problem just got bigger. After KPMG got caught, now it’s PwC. GPTZero analyzed four PwC Middle East reports published between 2024 and 2026 and found fabricated sources and false claims woven throughout. The worst case? A report called “Transforming Governance” that’s 84 percent likely to be entirely AI-generated.
PwC uses that specific report to pitch a product called “Citizen Pulse,” claiming governments in Denmark, Saudi Arabia, the U.S., and Australia already rely on it. GPTZero couldn’t find any evidence backing those claims. The pattern is clear: reports with the highest AI-generated scores also contained the most hallucinations. GPTZero’s team calls it “Vibe Citing” — references that have roughly the right shape but wrong titles, broken URLs, missing authors, and no actual connection to the claims they supposedly support. It’s the same phenomenon plaguing academic journals.
PwC Middle East told the Financial Times it takes accuracy seriously and is “updating a limited number of supporting citations.” That’s a carefully worded non-explanation. They didn’t address how fabricated government clients ended up in a published report, or why no human caught it before publication. GPTZero had previously flagged similar issues at KPMG, Deloitte, and Ernst & Young — meaning this isn’t one firm cutting corners. It’s systemic.
What’s really happening here is consultants treating generative AI as a first draft engine and skipping the verification step. The economics are obvious: reports that once took weeks of research and junior analyst time now get generated in hours. But when the citations are fake and the claims can’t be verified, the product being sold is essentially plausible-sounding fiction dressed up in corporate branding. The question nobody’s answering: if this is what’s in the public-facing marketing materials, what’s in the client deliverables?
