ChatGPT tax advice already costing UK firms, say accountants

UK accountants report that businesses are already losing money after using public artificial intelligence tools such as ChatGPT for tax, bookkeeping and financial advice, and warn that misuse could contribute to business failures as early as next year.

New research from Dext, based on a survey of 500 accountants and bookkeepers, found that half of respondents are aware of businesses that have suffered direct financial losses because of incorrect or misleading AI-generated advice. These losses include overpaid tax, missed allowances, penalties, fines and compliance issues.

The findings indicate that use of general-purpose AI in finance and tax decisions is growing across British businesses. Accountants say the trend is creating a widening risk as some firms treat AI outputs as if they were professional guidance.

"The damage is no longer hypothetical," said Paul Lodder, VP accounting product strategy at Dext. "Businesses are already losing money, and accountants are spending valuable time correcting avoidable mistakes, from VAT and payroll errors to misinterpretation of expenses.

"AI has a powerful role to play in finance but there's a fundamental difference between specialist tools built for accounting and bookkeeping, and general-purpose chatbots that don't know a business's true financial context."

Rising reliance

The survey suggests that client use of public AI tools has increased sharply during 2025. More than three quarters, 77%, of accountants and bookkeepers say they have seen more clients using systems such as ChatGPT and other large language models for financial, tax or bookkeeping queries.

Almost three in four, 72%, say more clients now use AI-generated outputs to question or challenge their professional advice. A further 68% report a rise in clients suggesting that AI could replace the need for professional accounting services.

Accountants link that growing reliance with a rise in errors that are already affecting client finances. Respondents describe a pattern of incorrect or incomplete guidance that is feeding directly into bookkeeping entries and tax filings.

Frequent errors

Nearly a third, 31%, of accountants and bookkeepers say they now encounter mistakes caused by wrong or misleading AI-generated financial or tax advice on a weekly basis. Seven per cent see such errors every day. A further 28% encounter them monthly. Only 5% say they have never seen a mistake linked to public AI tools.

The most common problems relate to classification of expenses and tax treatment. Forty-six per cent report incorrect interpretation of business expenses. Forty-one per cent have seen errors in claiming or charging VAT. Thirty-five per cent report flawed personal tax planning advice. Thirty-four per cent cite payroll errors, and another 34% point to incorrect business tax planning advice.

These mistakes can lead to misreported income and costs, underpayments or overpayments of tax, and inconsistencies that trigger intervention from authorities.

Time and cost

The research points to a growing workload for the profession as it addresses AI-related errors. Among those who encounter public AI-related mistakes, 93% estimate they spend up to ten hours each month correcting them.

Within that group, 44% spend up to three hours a month resolving issues created by AI-generated advice. Thirty-nine per cent spend between four and ten hours. That represents a material time burden for firms and a potential extra cost for their clients.

Accountants say the work involves unpicking incorrect entries, revising calculations and clarifying advice that clients previously obtained from AI tools.

2026 risks

Looking ahead, respondents express concern that the risks will grow if businesses continue to rely on unregulated AI tools without involving qualified professionals. A third, 33%, warn of a higher risk of insolvency or business failure in 2026 linked to misuse of AI outputs in financial decisions.

Forty-three per cent expect increased misuse of AI-generated content to justify inappropriate or fraudulent claims. Thirty-eight per cent foresee rising fines and penalties. Thirty-seven per cent anticipate greater scrutiny from HM Revenue & Customs due to incorrect or late filings. Nearly half, 45%, believe that business decisions based on false confidence in inaccurate AI outputs will become more common.

These concerns reflect a fear that mistakes will compound over time. Small errors in tax treatment or reporting may escalate into larger financial problems if they remain undetected across several periods.

Regulation calls

The survey shows broad support among accountants for tighter oversight of public AI tools when they touch financial matters. More than nine in ten respondents, 92%, say such tools should be regulated or restricted when providing financial or tax-related advice. Seventy per cent call specifically for formal regulation, while others suggest other forms of restriction.

Accountants argue that current systems do not distinguish between casual information requests and queries that have regulatory or legal consequences. They say clearer boundaries are needed around automated advice that could influence tax submissions, payroll and statutory reporting.

"If we head into 2026 with more businesses treating AI outputs as trusted tax and financial advice, without professional oversight, the consequences could be severe. The focus now should be on responsible guardrails, clearer restrictions around financial advice, and better education for businesses on what these tools can and cannot safely be used for," said Lodder.