KPMG withdraws AI report after UBS, NHS and Swiss Railways dispute claims

KPMG has withdrawn a report examining enterprise adoption of artificial intelligence after several organisations, including UBS, the UK National Health Service (NHS) and Swiss Federal Railways, said claims made about their use of AI were inaccurate or misleading.

The professional services firm removed the report, titled Redefining Excellence in the Age of Agentic AI, from its websites while conducting an internal investigation into the matter.

The move follows scrutiny of the report's contents after research group GPTZero identified a number of inaccuracies and suggested that some of the issues appeared consistent with AI hallucinations.

Organisations challenge statements in the report

According to reporting by the Financial Times, multiple organisations disputed how their AI activities were described in the publication.

Those cited by the newspaper included:

• UBS

• The UK National Health Service

• Swiss Federal Railways

• Transport for London (TfL)

The organisations told the Financial Times that claims regarding their AI usage were either untrue or misleading.

The report had focused on the growing role of agentic AI, a category of artificial intelligence systems designed to perform tasks with greater autonomy than traditional AI tools.

Questions emerged after GPTZero reviewed the publication and identified what it described as inaccuracies relating to some of the organisations featured in the report.

KPMG launches internal review

In response to the concerns, KPMG confirmed it had removed the report while examining the issues raised.

A spokesperson for the firm said the organisation expects employees to adhere to its policies governing the use of artificial intelligence.

"We expect all our people to follow our guidelines on the responsible use of AI, including human oversight to validate content and verify independent sources," the spokesperson said.

KPMG has not publicly detailed the outcome of its review or specified how the inaccuracies entered the report.

The incident has attracted attention because it involves a major professional services firm publishing research on AI adoption at a time when businesses are increasingly examining how to deploy AI responsibly.

Fresh scrutiny over AI-generated content

The controversy arrives amid broader concerns about the reliability of AI-generated material in corporate research, consulting and professional services.

AI hallucinations occur when artificial intelligence systems generate information that appears credible but is inaccurate, unsupported or entirely fabricated.

While many organisations use AI tools to assist with research, drafting and analysis, experts have repeatedly stressed the need for human verification before publication.

The KPMG case highlights the challenges companies face when integrating AI into knowledge-intensive work that relies on factual accuracy and source validation.

Not the first Big Four setback

The withdrawal also follows a similar incident involving another major professional services firm.

Last month, EY withdrew a report focused on loyalty rewards programmes after concerns emerged that it contained apparent AI hallucinations and fabricated footnotes.

The two incidents have intensified discussion around governance, quality controls and editorial oversight as professional services firms increasingly incorporate AI into internal workflows and client-facing research.

For firms that advise businesses on technology adoption and risk management, such episodes can carry reputational implications beyond the reports themselves.

A test of AI governance

As organisations accelerate investment in artificial intelligence, the KPMG episode serves as a reminder that AI-assisted content remains dependent on rigorous human review.

The outcome of the firm's investigation may provide further insight into how large consulting and professional services organisations are managing AI-related risks. More broadly, the incident is likely to reinforce calls for stronger verification processes as AI-generated content becomes more common in business research and corporate communications.