PwC has come under scrutiny after multiple reports produced by its Middle East division were found to contain fabricated references, misattributed claims, and unverifiable information, raising concerns about the reliability of AI-assisted content in professional consulting. The reports, compiled over the past two years, included analyses on artificial intelligence applications, electric vehicles, and government public service improvements.
An investigation by researchers at GPTZero revealed that four PwC Middle East documents exhibited signs of heavy reliance on AI-generated text, with numerous footnotes failing to substantiate the claims they supported. Some citations directed readers to nonexistent web pages or sources that did not contain the referenced material. Notably, one cited academic paper on air quality in Riyadh—purportedly published in a reputable journal and attributed to specific authors—could not be located in either the journal or the authors’ records, suggesting it was entirely fabricated by AI.
Other problematic references included an unusual citation of a teenage blogger with a small online following as the source for information about a JPMorgan initiative. The initiative, described as a “real world success story” of autonomous AI agents automating commercial loan agreement reviews, originally emerged in 2017, several years before generative AI tools like ChatGPT became widely available.
This incident echoes similar findings involving reports from other major consulting firms. Earlier investigations by GPTZero uncovered AI-generated errors in publications from Ernst & Young and KPMG, prompting those firms to retract certain documents.
PwC Middle East responded by stating it takes the accuracy of its published research seriously and is in the process of updating some citations. The firm emphasized its commitment to responsible AI practices and noted the existence of quality control procedures for research and content development. However, it did not provide details explaining how the errors were introduced into the reports initially.
The episode highlights the challenges consulting firms face in integrating AI tools into research and publication processes while maintaining rigorous standards of verification and accountability. As AI-generated content gains prominence, such lapses risk undermining trust in expert advice on emerging technologies and their ethical implementation.