KPMG Australia has fined one of its partners A$10,000 for using artificial intelligence (AI) to cheat on an internal AI training examination. This incident, which also saw the partner forced to retake the test, highlights a wider problem within the firm. More than two dozen other KPMG Australia staff have been caught using AI to cheat on internal exams since July, underscoring the challenges professional services firms face with new technology.^{[1][2][3][4][5][6][7][8][9]} ### Partner Used AI for AI Exam Answers The unnamed KPMG Australia partner, a registered company auditor, completed mandatory AI training in July. During the exam, the partner uploaded the recommended reference manual for the course into an AI tool. This AI tool then generated answers for the exam questions. KPMG's internal AI detection systems flagged the activity in August, leading to an investigation and the subsequent A$10,000 fine. The fine is equivalent to approximately $7,000 US dollars or £5,200 British pounds. The partner will also forfeit more than $10,000 in future income. ^{[1][2][3][5][6][7][8][9]} KPMG Australia CEO Andrew Yates acknowledged the difficulties posed by the rapid adoption of AI. "Like most organisations, we have been grappling with the role and use of AI as it relates to internal training and testing," Yates told The Australian Financial Review. "It's a very hard thing to get on top of, given how quickly society has embraced it." ^{[1][2][3][5][7][8][9]} ### Broader Cheating Concerns Emerge This partner's misconduct is not an isolated event. KPMG Australia confirmed that 28 staff members, including the partner and others at or below manager level, have been identified using AI to cheat on internal tests since July. The firm introduced AI monitoring in its internal testing in 2024. These cases highlight a persistent issue with exam integrity within the accounting sector. ^{[1][2][3][5][6][8][9]} KPMG has stated it will now publicly disclose AI-related cheating cases in its annual results. It will also verify that staff meet their obligations to self-report misconduct to relevant professional bodies. This move aims to increase transparency and pressure rival organizations to adopt similar standards. ^[1][5][8] ### Industry-Wide Struggle with AI Ethics The accounting industry has grappled with cheating scandals for several years. In 2021, KPMG Australia faced a A$615,000 fine over "widespread" misconduct. More than 1,100 partners were involved in "improper answer-sharing" on integrity and skill assessment tests. AI tools introduce new avenues for rule-breaking, adding another layer of complexity to these challenges. ^[2][3] The Association of Chartered Certified Accountants (ACCA), the world's largest accounting body, stopped remote exams late last year. The ACCA cited the difficulty of preventing AI-powered cheating with existing safeguards. Helen Brand, ACCA chief executive, said AI tools had created a "tipping point" where the use of AI was outpacing cheating prevention measures. ^[2][3][9] Another Big Four firm, Deloitte Australia, also faced issues related to AI use. Deloitte had to refund part of its fee for a government report that contained errors, including fabricated quotes and references to non-existent academic research papers, due to the use of generative AI. This incident led to a partner reportedly leaving the firm. ^[5] ### KPMG's Commitment to AI and Ethics Despite these challenges, KPMG actively promotes the adoption of AI within its operations and for clients. The firm offers various AI learning programs designed to equip leaders and employees with AI capabilities. These programs cover understanding AI's potential, risks, governance frameworks, and practical application. KPMG aims to help organizations embed AI into their operations and drive value. ^{[10][11][12][13]} KPMG also emphasizes the importance of AI ethics and governance. The firm has developed internal policies and a "Trusted AI framework" with ten ethical principles. These aim to guide responsible AI innovation and ensure data quality, integrity, and transparency in AI models. KPMG believes that ethical AI involves complex business, regulatory, and technical challenges. ^{[14][15][16][17]} Andrew Yates noted that KPMG has implemented new technologies to block access to AI during testing. The firm also conducted a significant firm-wide education campaign after first detecting instances of AI use outside its policy in 2024. ^[8] ### Moving Forward The partner's fine and the broader cheating statistics highlight the ongoing tension between embracing new technology and maintaining professional integrity. As AI becomes more integrated into daily work, professional services firms must constantly adapt their training, testing, and ethical frameworks. The incidents underscore the need for robust internal controls and clear policies to manage the risks associated with AI adoption. KPMG's decision to disclose AI-related cheating in its annual results sets a new standard for transparency in the industry. ^[1][5][8]

AI Insights Hub