Businesses across the United Kingdom are suffering tangible financial losses after relying on misleading advice from public artificial intelligence platforms, a major new study reveals. The research, published by accounting software provider Dext, indicates that the damage from so-called 'AI slop' is now very real, moving beyond theoretical risk into costly, avoidable mistakes.
The Scale of the AI Advice Problem
Nearly one in three (31 per cent) accountants and bookkeepers now encounter client mistakes on a weekly basis, errors directly traced to incorrect or misleading AI-generated guidance on finance and tax. This troubling frequency stems from a sharp increase in businesses turning to public chatbots for advice before consulting a professional, a trend revealed last October which showed almost two-thirds of firms engage in this practice.
Paul Lodder, Vice President of Accounting Product Strategy at Dext, stated bluntly: "The damage is no longer hypothetical." His warning underscores a shift from concern about potential pitfalls to the documentation of actual harm. The research found that half of all UK accountants and bookkeepers have seen businesses suffer direct financial losses due to these AI-induced errors.
Most Common and Costly Errors
The study pinpointed the specific areas where faulty AI advice is hitting businesses hardest. The most frequently reported mistakes include:
- Incorrect interpretation of business expenses (46 per cent)
- Wrong VAT claims or charges (41 per cent)
- Flawed personal tax planning (35 per cent)
- Payroll errors (34 per cent)
- Incorrect business tax planning advice (34 per cent)
Beyond the immediate financial hit, these errors are creating a secondary cost burden. Accountants are spending increasing amounts of billable hours hunting down and rectifying mistakes made by AI, driving up fees for the very businesses trying to save money.
HMRC Crackdown and Rising Risks for 2026
Accountants are warning that the risks will intensify significantly in 2026 if current trends continue. Over 30 per cent of practitioners have warned of a higher risk of insolvency or outright business failure for clients who persist in using AI tools without professional oversight.
Other anticipated consequences include increased misuse of AI outputs to justify inappropriate or even fraudulent claims (43 per cent), a rise in fines and penalties (38 per cent), and greater scrutiny from HM Revenue & Customs due to incorrect or late filings (37 per cent). This heightened HMRC attention is not an empty threat. The tax authority plans to recruit 5,000 new compliance officers by 2029/30, with the hiring drive set to begin in 2026, signalling a major ramp-up in enforcement actions.
Lodder emphasised the gravity of the situation: "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." In response to these dangers, an overwhelming majority—more than nine in ten accountants—believe public AI tools should be subject to regulation, or even restriction, when providing financial or tax-related guidance.
