A KPMG report on AI in business included fabricated case studies, according to a recent investigation. The report, titled 'Redefining excellence in the age of agentic AI,' made false claims about AI implementation at UBS, the UK's NHS, Swiss Federal Railways, and Transport for London. GPTZero identified these inaccuracies, and the Financial Times confirmed the findings. All named organizations denied the claims, highlighting the report's unreliability. The incident has raised concerns about the credibility of AI-driven consulting reports.
GPTZero CEO Edward Tian warned that flawed reports from major consulting firms can lead to 'secondary hallucinations,' as these reports are often considered highly credible and reused by both AI systems and people. The firm also highlighted issues with sourcing, which may have contributed to the errors. GPTZero noted that citations were often loose paraphrases of real sources, missing URLs or correct authors. In some cases, no original source existed at all. This practice, termed 'vibe citing,' is also a problem with Google's AI Overviews, which a German court recently ruled Google liable for.
KPMG has removed the report from several websites. The incident is particularly damaging for the firm, as it spread misinformation and demonstrated an inability to manage the very technology it aims to sell to its clients. Source: thedecoder
