When auditors discuss AI, the debate often moves quickly to technology. Which tool is being used? How powerful is the model? Can the output be trusted? These are valid questions. But a recent LinkedIn poll I conducted suggests that professionals see a deeper concern. The greatest threat to auditor independence may not be AI itself. It may be the auditor’s overconfidence in AI-generated answers.
The opinion poll attracted more than 170 responses and more than 20 qualitative comments. A striking 76% said that over-reliance on AI-generated outputs is the real issue that may affect auditor independence. A further 12% pointed to opaque AI models as the key risk. Only 7% selected commercial pressure on audit firms, while 4% chose conflicts of interest and relationships.
These results are telling. Traditional independence concerns have not disappeared, but the profession appears more worried about a new behavioural risk: the quiet transfer of judgment from auditor to algorithm.
Understanding independence
This is where independence needs to be understood more broadly. Independence is not only about financial interests, commercial relationships or familiarity threats. It is also about independence of mind. If an audit team accepts an AI-generated conclusion without sufficient challenge, the outward appearance of independence may remain intact, but professional scepticism may have weakened. The auditor may still sign the report, but the thinking behind that signature may have been outsourced.
AI outputs often sound confident, structured and technically persuasive
That risk is subtle because AI outputs often sound confident, structured and technically persuasive. This is precisely why over-reliance matters. An answer can be fluent and still be wrong. A summary can be neat and still omit a critical exception. A risk assessment can look complete and still miss the one matter that should have called for deeper enquiry. In audit, the danger is rarely a tool that openly fails; it is a tool that appears convincing enough to reduce the auditor’s instinct to challenge.
More than opacity
Research suggests that AI can support audit quality when used carefully, but that audit judgment must remain with the engagement team. But there are ongoing concerns about the governance and transparency challenges created by black-box models, where the inputs and outputs may be visible but the reasoning in between is not. That aligns closely with the 13% of respondents who selected opaque AI models as the main concern.
Yet the poll result suggests that opacity alone is not the profession’s primary anxiety. Auditors seem to recognise that opacity becomes dangerous when combined with dependence. A black-box tool is a governance issue; a black-box tool that the auditor stops questioning is an independence issue.
AI should sharpen the auditor’s mind, not replace it
The lower scores for commercial pressure and conflicts of interest should not be misread. These remain core threats under established ethical frameworks and will continue to demand safeguards.
However, the poll indicates that AI introduces a different pressure point. It may not come from a client relationship or fee dependency; it may come from speed, convenience and the institutional desire to do more with less. If AI shortens procedures, reduces friction and presents ready-made conclusions, audit teams may face a new kind of pressure: the pressure not to slow down and think.
Govern with discipline
The response for firms is not to resist AI. That would be unrealistic and, in many cases, unhelpful. AI can enhance audit quality by improving data analysis, identifying unusual patterns, summarising large volumes of information and helping teams ask better questions. But it must be governed with discipline.
Firms need clear policies on approved tools, documentation of AI-assisted work, review protocols, escalation procedures and training that reinforces professional scepticism. The test should not be whether AI was used. The test should be whether the auditor can explain, evidence and defend the conclusion independently.
For regulators, standard setters and firms, the message is clear. The future of auditor independence will not be protected by disclosure alone. It will be protected by judgment, accountability and governance. AI should sharpen the auditor’s mind, not replace it.
If the profession gets this balance right, AI can become a force for stronger assurance. If it gets it wrong, independence may be weakened – not by obvious compromise but by quiet reliance.
More information
Find out more about the use of AI in the workplace at ACCA’s Accounting for the Future conference, including the session ‘Evolving roles in a data-driven age’. Register to watch live on 24-26 November or on demand
