A partner at KPMG was fined A$10,000 (£5,200) for using artificial intelligence to cheat during an internal AI training course. This is part of a growing problem, with more than two dozen KPMG Australia employees caught using AI tools to cheat on exams since July. The company used its own AI detection software to find the cheating, according to the Australian Finance Review. KPMG is not new to cheating issues; in 2021, it was fined A$615,000 for improper answer-sharing involving over 1,100 partners. AI has now opened new doors for rule-breaking. In December, the UK's Association of Chartered Certified Accountants announced that students must take exams in person to prevent AI cheating. Helen Brand, ACCA chief executive, called AI use a “tipping point” surpassing anti-cheating safeguards. Meanwhile, KPMG and other firms like PricewaterhouseCoopers encourage staff to use AI to improve profits and reduce costs. KPMG plans to judge partners on AI skills in 2026 reviews. Niale Cleobury, KPMG’s global AI workforce lead, said, “We all have a responsibility to be bringing AI to all of our work.” Some voices online see the cheating as a training problem, not just a cheating one. KPMG’s CEO Andrew Yates admitted managing AI’s impact on training is tough. He said, “Given the everyday use of these tools, some people breach our policy. We take it seriously when they do.” KPMG promises to strengthen its approach and monitor AI misuse among staff.