Updated
Updated · Accounting Times · Jun 18
Australian Firms Lose Money From AI Tax Advice as 92% of Accountants Back Regulation
Updated
Updated · Accounting Times · Jun 18

Australian Firms Lose Money From AI Tax Advice as 92% of Accountants Back Regulation

2 articles · Updated · Accounting Times · Jun 18

Summary

  • 63% of 500 Australian accountants and bookkeepers surveyed said they knew businesses hit by direct losses after using general-purpose AI for tax, bookkeeping or financial advice.
  • 75% said client use of public AI tools such as ChatGPT and Claude is rising, while the errors they are correcting include overpayments, missed allowances, payroll mistakes, penalties and compliance breaches.
  • 92% backed intervention, including 69% calling for formal regulation and 48% supporting restrictions on public AI tools giving financial or tax advice.
  • Dext said small businesses often turn to AI to cut advisory costs, but accountants warned the mistakes can consume professional time, trigger regulatory scrutiny and in severe cases raise insolvency risks.

Insights

Is AI's 'free' financial advice costing businesses more than professional help through hidden errors and compliance penalties?
Who is legally responsible when AI gives bad financial advice—the user, the developer, or no one at all?

The Rising Cost of AI Errors: Financial and Reputational Risks Facing Australian Businesses in 2025-2026

Overview

In the 2025-2026 financial year, Australian businesses have faced rising financial and reputational costs due to errors and misuse of AI, especially as adoption increases. As more companies rely on AI, the risk of critical mistakes grows, leading to what is described as a 'hallucination tax'—ongoing losses from damaged trust, regulatory penalties, and customer harm. Public concern is also growing, with nearly a third of surveyed Australians and New Zealanders worried about AI’s accuracy. High-profile incidents, like a KPMG partner being fined for improper AI use, highlight the real financial and professional consequences of relying on AI without proper oversight.

...