2 hours ago
5
ARTICLE AD BOX
KPMG has pulled its planetary study connected “Agentic AI” aft large companies and organisations named successful it complained that their expected achievements were wholly made up. The nonrecreational services elephantine published the report, titled “Redefining excellence successful the property of agentic AI,” with galore mendacious claims and bogus lawsuit studies, according to a study by The Financial Times. The companies says the fabricated occurrence stories look to person been generated by artificial quality (AI) “hallucinations” – wherever an AI exemplary confidently includes incorrect facts successful responses which KPMG’s quality unit failed to catch.The inaccuracies were archetypal discovered by the tech probe radical GPTZero and aboriginal verified by the Financial Times. After being alerted to the fake data, respective high-profile organisations, including Swiss slope UBS, the UK's National Health Service (NHS) and large nationalist transit agencies, forced KPMG to region the work from its websites.
What the KPMG study claimed and what companies said
KPMG’s study claimed however planetary organisations were successfully deploying highly precocious “AI agents” to grip complex, automated tasks.
However, the organisations progressive rapidly clarified.UBS Bank: The study claimed the planetary wealthiness manager integrated AI agents crossed its concern advisory and hazard absorption systems utilizing a customized level built with Microsoft. A spokesperson for UBS reportedly told reporters the claims were “factually incorrect” and demanded their removal.Swiss national railways: KPMG wrote that the railway utilized AI agents to assistance users seamlessly program and publication journeys based connected real-time conditions and c footprints.
A railway spokesperson confirmed the claims were “not accurate”.Transport for London (TfL): The survey claimed London’s transit strategy was utilizing AI agents to foretell congestion and coordinate metropolis transport – a assertion dubbed “misleading” by a TfL spokesperson.NHS Greater Manchester: The study claimed the wellness work utilized AI agents to organise diligent data, automate referrals and foretell infirmary readmissions.
A spokesperson revealed the assertion “doesn't truly align” with reality.
‘Poisoning the Well’ of trust
The KPMG ungraded is 2nd large event. Just past month, rival steadfast EY was forced to retract a large survey aft GPTZero caught fake footnotes and different AI-generated errors successful its text. As the “Big Four” accounting and consulting firms are typically viewed arsenic highly credible, their nonaccomplishment to fact-check their ain enactment creates a unsafe ripple effect."They poison the good of information," said Edward Tian, main enforcement of GPTZero. Tian pointed retired that earlier KPMG pulled the report, its fake findings had already been cited by aggregate tech manufacture publications and a large European newspaper.
KPMG launches investigation
A spokesperson for KPMG International stated that the steadfast takes the “accuracy and integrity of its published contented seriously” and confirmed the study has been removed portion a afloat interior probe takes place.The steadfast admitted that its ain employees apt broke interior rules regarding artificial intelligence. “We expect each our radical to travel our guidelines connected the liable usage of AI, including quality oversight to validate contented and verify autarkic sources,” the spokesperson was quoted arsenic saying.
