KPMG Australia Partner Reportedly Used AI to Cheat on Internal AI Training Test and Was Fined A$10,000

In Australia, a KPMG Australia partner and registered company auditor reportedly uploaded a reference document from an internal AI training course into an AI tool to answer an exam question, in violation of firm policy. KPMG reportedly detected the activity in August 2025 and imposed a penalty of more than A$10,000 of future income after an internal investigation. The partner also reportedly self-reported the matter to Chartered Accountants ANZ, which is investigating the case.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation

KPMG employees used AI tools to cheat on internal training exams to gain an unfair personal advantage, which fits under AI-facilitated cheating and plagiarism.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by human employees intentionally using AI tools to cheat on internal exams after the AI tools were deployed.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes the use of general-purpose AI tools to answer exam questions, which falls under the category of chatbots or general AI-generated content.

AI system and alleged parties

• AI system: AI tool
• AI purpose: Question Answering; Writing Assistant
• Behaviour type: Tool
• Alleged developer: Unknown generative AI developers
• Alleged deployer: Unnamed KPMG Australia partner
• Alleged harmed parties: KPMG Australia

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Minor, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report explicitly describes financial losses in the form of a fine imposed on a partner. The total reported loss is a fine of more than $10,000 for 1 partner across 28 occurrences, averaging approximately $357 per occurrence.

Directly caused: The partner was fined more than $10,000 of future income as a direct penalty for the cheating incident.

People affected

• Occurrences reported: 28
• People reportedly harmed: 1
• People reportedly exposed: 28

Potential causes

Management

• Slow Safeguard Implementation: Management fell behind rapid AI adoption before introducing monitoring.
• Push for AI Adoption: Prioritizing AI integration without setting clear operational boundaries.

Technology

• Easy Access to AI Tools: Rapid societal adoption of AI tools made them easily accessible for cheating.
• Lack of AI Blocking Tools: Initial lack of technical blocks allowed access to AI tools during exams.

Data Inputs

• Uploadable Reference Manuals: Course materials were downloadable, enabling users to upload them into AI.

Human Factors

• Unethical Behavior by Staff: Personnel, including a partner, chose to violate policy to pass exams.
• Pressure to Use AI: Mandates to use AI at work created confusion over acceptable boundaries.

Process and Methods

• Outdated Training Design: Testing methods did not adapt to the existence of AI capabilities.
• Open-Book Testing Vulnerability: Open-book format allowed easy copy-pasting or uploading of text to AI.

Regulatory Environment

• Weak Reporting Requirements: Firms are not obligated to report exam misconduct to ASIC immediately.
• Reliance on Self-Reporting: The regulatory regime relies on individuals self-reporting ethical breaches.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear and consistent details about the number of cheating cases (28), the specific actions taken (uploading reference materials to AI tools), and the resulting disciplinary actions (a partner fined $10,000). There is little ambiguity about the events.
• Ambiguities identified: The specific AI models used by the employees are not named, referred to only generally as AI tools.

The incident involves 28 KPMG Australia personnel using AI tools to cheat on internal training exams, leading to internal disciplinary actions and political discussions regarding corporate self-regulation. It represents a negligible national security threat, categorized primarily as a corporate compliance and professional ethics issue rather than a threat to state security.

• Overall national security impact: Negligible
• Scope: Single nation
• Primary target: Australia
• Alleged perpetrator: KPMG Australia personnel

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic concern regarding the governance of AI use in professional and educational settings.
• Autonomy: Human-controlled. The AI tools were directly used by humans as assistants to generate exam answers based on uploaded materials.
• Novelty: Established threat. Using AI tools to cheat on exams is an established behavior, though its occurrence within a major consultancy firm drew regulatory attention.

Impact by dimension

• Physical security: Negligible. No physical systems, infrastructure, or human safety was threatened by this corporate training cheating incident.
• Information security: Negligible. No compromise of intelligence capabilities, classified models, or systematic information manipulation occurred.
• Sovereignty: Negligible. The incident sparked local political debate on corporate regulation, but did not threaten state authority, elections, or core government operations.
• Economic security: Negligible. While involving a major accounting firm, the incident represents internal professional misconduct rather than a systemic threat to economic stability or strategic technology.
• Societal stability: Negligible. No threats to social cohesion, civil liberties, or population safety were identified in this corporate misconduct case.