Keeping the judgment in AI audits

Multiple-choice questions. In order to be awarded CPD units you must answer the following five random questions correctly. If you fail the test, please reread the article before attempting the questions again.

What is a key concern about AI-generated output in audit?

It can appear credible even when it is wrong or misleading

The skills needed to apply AI are not yet widespread

The data it is based on might be flawed

It is often inaccurate

How can AI add value to an audit?

By speeding up audit processes

By helping identify inconsistencies, unusual items and areas requiring further enquiry

By offering reassurance to the client

By increasing the chance of audit accuracy

What is ‘omission risk’?

AI relying on outdated data

AI summarising information but leaving out important details

AI producing duplicate reports

AI bypassing standard protocols

Which of the following is an example of interpretation risk?

AI misunderstanding the human auditor’s instruction

AI misunderstanding the commercial substance of a contract

AI failing to understand/comply with audit standards

Tha human auditor giving AI incorrect commands

What gives clients confidence in AI-assisted audits?

Knowing AI is used within a framework of evidence, scepticism, accountability and review

Knowing the audit firm uses the latest AI tools

Receiving more accurate audit reports

Receiving more detailed audit reports

The author suggests that the value of AI in audit should primarily be assessed based on its ability to:

Automate the majority of audit procedures

Reduce resourcing requirements

Enhance audit quality by supporting better risk identification and enquiry

Improve efficiency

Which example demonstrates AI's potential to improve risk assessment?

Being able to filter out non-material risks

Automatically identifying risks not picked up by human auditors

Identifying inconsistencies between board minutes and management's going concern assessment

Calculating materiality thresholds independently

Why does the author argue that AI cannot replace the auditor's role?

AI use in audit is not permitted by all regulators globally

AI is currently unable to interface with auditing standards

AI cannot perform substantive testing

AI does not exercise professional scepticism, understand client-specific context or assume responsibility for conclusions

Which of the following would be classified as a hallucination risk?

AI incorrectly interprets management's explanation of a transaction

AI analyses outdated financial information

AI invents a reference to supporting evidence that does not exist

AI omits a restrictive covenant from a loan agreement summary

Which situation best illustrates methodology risk?

Failing to document AI usage in the audit file

Using an outdated version of an AI model

Entering confidential client information into an unapproved AI tool

Relying on AI-generated procedures without evaluating whether they provide sufficient appropriate audit evidence

1 Unit