Global consulting giant PwC has come under scrutiny after several of its “thought leadership” reports were found to contain AI-generated inaccuracies, including questionable references, incorrect citations and claims that could not be verified.
The reports, produced by PwC’s Middle East division, focused on emerging technology topics including artificial intelligence, autonomous vehicles and electric vehicles. An investigation by AI research group GPTZero, reviewed by the Financial Times, found that some documents appeared to rely heavily on generative AI tools without sufficient human verification.
Researchers identified problems including fabricated academic references, incorrect source links and misattributed information. Some footnotes reportedly referred to sources that did not exist, raising concerns about the reliability of research published by one of the world’s largest professional services firms.
The controversy is particularly significant because PwC advises companies and governments on responsible AI adoption and risk management. Experts say the incident highlights a wider challenge facing organisations using AI: productivity gains must be balanced with strong review processes and human oversight.
PwC said accuracy remains a priority and that it has quality control procedures for published research. The firm indicated that some supporting citations would be updated following the findings.
The issue follows similar cases involving other major consulting firms, with investigations previously identifying AI-generated errors in reports from firms including EY and KPMG. The incidents have intensified debate over how professional organisations should manage AI tools when producing public research and advisory material.
The controversy serves as a warning for businesses adopting generative AI: while the technology can accelerate research and content creation, unchecked AI outputs can introduce serious reputational risks if they are not carefully reviewed.
