PwC Middle East Under Fire as Investigation Uncovers AI-Generated Claims and Hallucinated Citations in Research Reports
Global accounting firm PwC is facing scrutiny after an independent investigation revealed that several research publications from its Middle East arm contained fabricated statistics, ghost citations, and hallmarks of unverified AI content generation.
An investigation led by AI detection firm GPTZero analyzed four “thought leadership” papers published by PwC Middle East between 2024 and 2026. The probe concluded that the materials relied heavily on unchecked artificial intelligence outputs, with one report identified as entirely machine-generated.
The Findings: Unverified Claims and Ghost Citations
According to GPTZero analysts Paul Esau, Om Ogale, and Alex Cui, the firm’s 2025 publication titled ‘Transforming Governance’ showed an 84% likelihood of being AI-generated—a figure that rose to 100% when excluding the references section.
The report promoted a proprietary PwC data framework called Citizen Pulse, claiming active deployments by public sector authorities in the United States, Denmark, Saudi Arabia, and Australia. However, researchers found no public record or verifiable evidence that any such government contracts existed.
GPTZero researchers highlighted several red flags typical of Large Language Model (LLM) outputs:
- Mismatched References: Only 10 of the 17 listed sources in the bibliography matched actual footnotes in the body text.
- Inconsistent Terminology: The text fluctuated between describing Citizen Pulse as a conceptual framework and presenting it as a commercially deployed product—a signature issue caused by poor AI prompting.
- Direct AI Links: In a separate report titled ‘Building a Cyber-Resilient eMobility Ecosystem’, investigators discovered a source URL pointing directly to a ChatGPT session link.
- Flawed Data: Other analyzed reports included conflated market metrics, mismatched source citations, and false statistical correlations.
PwC’s Response
In response to the inquiry, a spokesperson for PwC Middle East stated:
“PwC Middle East takes the accuracy of our published research seriously and is updating a limited number of supporting citations in the reports mentioned. Consistent with our approach to Responsible AI, we have quality control processes for research and content development we expect all our people to adhere to.”
The episode underscores growing concerns across the management consulting sector regarding the unchecked reliance on generative AI for corporate research and client-facing thought leadership.
Source: Originally reported by City A.M.