PricewaterhouseCoopers is involved in the "AI fraud" controversy in its research report, which is full of errors and omissions and has attracted public attention. PricewaterhouseCoopers (PwC), one of the world's four largest accounting firms, has been involved in a whirlpool of public opinion for allegedly publishing inferior content generated by AI as an authoritative research report.

PricewaterhouseCoopers is involved in the "AI fraud" controversy in its research report, which is full of errors and omissions and has attracted public attention. PricewaterhouseCoopers (PwC), one of the world's four largest accounting firms, has been caught in a whirlpool of public opinion for allegedly publishing inferior content generated by AI as an authoritative research report. According to investigations by third-party AI detection tool GPTZero and multiple mainstream media, at least four "thought leadership" special reports released by PwC's Middle East branch between 2024 and 2026 are filled with a large number of AI illusions, forged quotes and false footnotes. These reports, which were originally intended to showcase PwC’s insights into cutting-edge areas such as generative AI, public services and electric vehicles in the Middle East, exposed many ironic loopholes in the details. In a report claiming that "nearly 70% of regional CEOs believe that generative AI will reshape business," the footnote link it cited pointed to a webpage that did not mention any relevant investigations at all. The reference link of another cybersecurity report even directly carried the traceability label "utm_source= ChatGPT.com", which almost confirmed that the content was written by ChatGPT. What’s even more surprising is that when the report cited a “successful case” of JPMorgan Chase’s use of AI to automate the review of commercial loan agreements, the only source cited was a personal blog posted on Medium by a teenager with only 280 followers, and the report mistakenly cited an incident that occurred in 2017 as a project driven by large language models in recent years. In addition, some electric vehicle reports also cite untraceable research reports, footnotes point to invalid web pages, and repeat the same data three times in just two pages but use different sources or do not indicate the source at all. In this regard, policy analysts at GPTZero pointed out that this kind of confusing and illogical document annotation is a typical feature of AI-generated research. It is particularly ironic that PwC has always portrayed itself as an industry authority on "responsible use of AI" and now makes such low-level mistakes. In fact, PwC is not alone. Its competitors EY and KPMG have also previously been canceled and retracted related articles for publishing reports containing AI hallucinations. At present, PwC Middle East has admitted its mistake, stated that it attaches great importance to the accuracy of published research, and promised to update and correct the affected citations, and adjust relevant review policies to prevent such incidents from happening again. via cnBeta.COM - Chinese industry information station (author: source: cnBeta.COM)