When AI gets it wrong: Responsible use of AI in accountancy firms

Recent incidents involving a Big Four accountancy firm in Canada and Australia have again highlighted the risks of reliance on Generative AI tools without proper safeguards, serving as a warning sign for UK accountancy firms as new technology provisions come into play under the ICAEW Code of Ethics. 
 

The recent Deloitte AI mishaps 
 

In October 2025, fabricated quotes and citations were discovered in an assurance review prepared by Deloitte Australia for the Australian Department of Employment Workplace Relations. The culprit? AI-generated hallucinations. It was confirmed shortly after that a Generative AI tool had been used in drafting the report. The fallout included reputational damage and a partial refund of fees around AUD$290,000 (£225,000). 

More recently in November 2025, Deloitte Canada has been alleged to have prepared an approximately CAD$1.6 million (£860,000) report for the Canadian Department of Health and Community Services containing potentially AI-generated errors. A spokesperson for Deloitte Canada denied the use of AI to write the report, but noted that a small number of corrections to the citations in the report would be made, acknowledging that AI was used selectively in respect of some citations.

Both recent mishaps highlight the significant financial and reputational risks of deploying AI tools without robust oversight. With the FRC’s guidance on this topic and the ICAEW’s new technology provisions in the Code of Ethics, UK regulators are clearly paying close attention. 

The UK regulatory backdrop 
 

The Financial Reporting Council (FRC) issued landmark guidance in June 2025 on AI use in audit. This guidance was published alongside a thematic review on certification processes of Automated Tools and Techniques (ATTs) in audits. The findings of the thematic review were stark, including that the six largest audit firms in the UK lacked up-to-date certification processes and monitoring capabilities for AI-driven tools. 

Around this time, in July 2025, the ICAEW also updated its Code of Ethics. The update included new sections on professional competence, confidentiality, and managing ethical threats arising from the use of technology. While these new Code of Ethics provisions remain largely untested as yet, any ICAEW member or member firm associated with a report containing AI-generated hallucinations or false citations could potentially expect scrutiny by the regulator.

Key ethical questions for accountancy firms
 

Under the new ICAEW Code of Ethics, member firms should consider the following matters: 

1. Is the use of the technology appropriate for its intended purpose? 
   In assessing whether the use of the technological output is appropriate for its intended purpose under R320.11, firms should consider the extent to which reliance will be placed on the output, the appropriateness of testing and evaluation processes, the extent of oversight in the technology’s design and development, as well as any controls regarding use of the technology.

2. Are there any self-interest threats? 
   R300.6 A2 warns against the use of technology where accountants themselves might not have sufficient information or expertise to explain it. Factors such as the extent of corporate oversight, internal controls and training processes therefore all become relevant when assessing the level of threat posed.  

3. Are the ICAEW’s fundamental principles upheld? 
   R113.1 and R112.1 highlight that professional judgment, objectivity, and competence must not be undermined by undue reliance on technology. Accountants should continue to foster their awareness and understanding of technological developments, being mindful of the principle of professional competence and due care.

Looking ahead 
 

In light of ever-evolving technological developments, firms would be well-advised to review their AI governance policies to maintain compliance with regulatory guidance.   

While the FRC guidance on AI use relates to the audit profession specifically, it nonetheless offers a useful starting point for firms considering development or use of AI tools. Firms should have at the forefront of their mind answers to questions such as: 

• What is the tool and how will it be used? 
• What criteria must be met before it is deployed? 
• How robust were the development and testing processes?
• What training and support are available for users of the tool? 
• Does it align with the UK government’s five AI principles? 

Likewise, the ICAEW has also emphasised in recently published guidance that firms should carefully examine risks and ethical considerations before commencing use of AI. In particular, firms are advised to implement clear policies and guidelines on AI use, ensure adequate staff training, maintain data quality, and continually challenge AI outputs with professional scepticism. 

Conclusion  
 

Accountancy firms are increasingly adopting AI for efficiency gains, but ethical and responsible use remains non-negotiable. The Deloitte cases serve as a stark reminder that without proper guardrails, firms risk financial loss, reputational damage, and regulatory consequences for improper AI use. Professional rigour and robust oversight must remain at the heart of any AI strategy and rollout.

For more on this topic please see Ian’s article published in Business and Accountancy Daily in November 2025. 

About the author

Ian Ko is a Senior Associate in the Regulatory team at Kingsley Napley LLP, qualified in New Zealand and New York. He specialises in advising professional services firms and individuals who are subject to regulatory investigations and enforcement proceedings, particularly in the accountancy and audit sector, as well as those seeking advice about regulatory compliance, including on the ethical use of AI in their organisations. 

Sam Binymin is a Senior Paralegal in the Regulatory team and also contributed to this article.