This article is for informational purposes only and does not constitute audit or legal advice. You should consult with a qualified professional before making any decisions about the use of AI in your firm.

---

How Can I Maintain Professional Scepticism When Using AI in Audit?

Professional scepticism is the bedrock of the audit profession. It is the questioning mind, the critical assessment of evidence. But the rise of artificial intelligence presents the most significant challenge to this fundamental principle in a generation. The allure of AI's speed and apparent certainty threatens to erode the very scepticism that gives the audit its value. And the FRC is watching.

ISA (UK) 200 is clear: the auditor must plan and perform an audit with professional scepticism. This is not optional. The FRC's recent guidance on AI in audit reinforces this, stating that "distance or digital mediation does not reduce the need for professional scepticism." In other words, just because the answer came from a sophisticated algorithm does not mean it is right.

The Seduction of the Black Box

AI tools can analyse millions of transactions in seconds, identify anomalies that a human would miss, and present their findings in a clean, confident dashboard. The temptation is to accept these findings at face value. The tool is unbiased, it is data-driven, it must be correct. This is a dangerous fallacy.

This is what I call the "seduction of the black box." The inner workings of the AI are opaque, so we defer to its judgment. We stop questioning, we stop challenging, and we stop thinking critically. We stop being auditors.

Why AI Demands More Scepticism, Not Less

Contrary to what some vendors might tell you, AI does not eliminate the need for professional scepticism; it amplifies it. You are no longer just auditing the client's financial statements; you are also, in effect, auditing the AI tool itself.

Here is why AI demands a heightened level of scepticism:

AI Risk	Implication for Professional Scepticism
Algorithmic Bias	The AI may have been trained on biased data, leading it to systematically ignore certain types of fraud or misstatement. You must question whether the AI's view of "normal" is appropriate for this specific client.
Model Hallucination	Generative AI tools can, and do, invent facts, references, and summaries. You must treat every piece of AI-generated text or analysis as an assertion to be verified, not as evidence in itself.
Overfitting & Lack of Context	The AI may be brilliant at identifying patterns in the data it was trained on, but it lacks the real-world, client-specific context that you, the human auditor, possess. You must question whether the anomalies it flags are genuine risks or just statistical quirks.
Data Integrity	The principle of "garbage in, garbage out" has never been more relevant. If the data fed into the AI is incomplete or inaccurate, its outputs will be worthless. You must be sceptical about the quality of the data before you even consider the AI's findings.

A Practical Guide to Sceptical AI Auditing

So, how do you maintain professional scepticism in the age of AI? It requires a conscious and deliberate shift in mindset. It is about treating the AI as a new, incredibly powerful, but fundamentally unreliable junior member of your audit team.

Here are five practical steps:

1. Question the Premise: Before you even run the tool, question its suitability for the task. Is this the right tool for this specific client and this specific audit assertion? What are its known limitations?
2. Challenge the Inputs: Do not just accept the data population. Perform your own checks to ensure the data fed into the AI is complete and accurate. Reconcile it to the trial balance.
3. Interrogate the Outputs: Never accept the AI's output at face value. Treat it as a starting point for your own investigation. If the AI flags 100 risky transactions, do not just sample 10 of them. First, try to understand why the AI flagged them. What are the common characteristics? Does it make sense in the context of the business?
4. Seek Disconfirming Evidence: This is the heart of professional scepticism. Do not just look for evidence that confirms the AI's findings. Actively look for evidence that contradicts them. If the AI says there are no anomalies in revenue recognition, your job is to go and find one.
5. Document Your Scepticism: Your audit file must tell the story of your sceptical mindset. It must show how you challenged the AI, how you verified its outputs, and how you reached your own independent conclusion. Your documentation should answer the question: "How did I satisfy myself that the AI's output was reliable?"

The Bottom Line: AI is a Tool, Not a Substitute for Judgment

AI has the potential to revolutionise the audit, to make it more efficient, more effective, and more insightful. But it will only achieve that potential if it is wielded by auditors who have not surrendered their professional judgment.

The FRC is not anti-AI. It is anti-complacency. It is anti-auditors who blindly follow the output of a black box.

Maintaining professional scepticism in the age of AI is not just a matter of regulatory compliance. It is a matter of professional survival. It is the thing that will continue to make you, the human auditor, indispensable.

---

Take the Next Step

If you are ready to move from theory to action, I can help. My AI Audit gives you a comprehensive assessment of your firm's AI readiness, identifying the gaps in your governance, the risks in your current tooling, and a clear roadmap to get you where you need to be.

Book a Discovery Call → or learn more about the AI Audit.