PwC faces scrutiny over AI-linked reports with fake citations and broken references

MUMBAI: It is a case of the footnotes stealing the spotlight. PwC is facing mounting scrutiny after several of its Middle East-focused research reports were found to contain fake citations, broken links and references to studies that investigators could not verify, raising fresh concerns over the growing use of artificial intelligence in professional research.

The issue came to light following an investigation by GPTZero, alongside reporting by the Financial Times. The reports in question, published between 2024 and 2026, covered subjects ranging from artificial intelligence and cybersecurity to governance and electric vehicles.

According to the investigation, numerous references either directed readers to non-existent webpages, cited sources that did not support the claims being made or referred to research papers that could not be located.

Among the most notable examples was a report referencing an academic study on air quality in Riyadh. Investigators said they were unable to find any evidence that the paper had been published in the journal cited or authored by the researchers named, suggesting the reference may have been fabricated.

GPTZero examined four PwC Middle East reports and concluded that they showed signs of extensive AI-assisted drafting. Researchers also identified what they described as “hallucinated citations” or “vibe citations”, references that appear credible at first glance but fail to substantiate the statements they accompany.

The findings have drawn particular attention because PwC advises organisations on artificial intelligence adoption, governance and risk management. The incident has renewed debate over whether companies are relying too heavily on generative AI without adequate human verification, particularly for thought leadership and research publications.

Responding to the findings, PwC Middle East told the Financial Times that it takes the accuracy of its research seriously and is updating a limited number of supporting citations in the affected reports. The firm said it has quality-control procedures in place for research and content development but did not detail how the problematic references passed internal review.

The controversy is not an isolated one. Earlier investigations by GPTZero led to EY and KPMG withdrawing reports that allegedly contained AI-generated inaccuracies, while Deloitte also came under scrutiny over documents that included fabricated references.

As consulting firms increasingly embrace artificial intelligence to accelerate research and content creation, the PwC episode serves as a reminder that while AI can help draft reports, credibility still depends on careful human review. In the race to adopt AI, it appears that checking the fine print remains just as important as writing it.