A&A Playbook - Virtual-2 - Sarb - ntersection-of-AI-Use-and-Audit-Documentation-Standards

2 - Sarb - ntersection-of-AI-Use-and-Audit-Documentation-Standards

Pdf Summary

The course explains how artificial intelligence affects audit work and documentation requirements. AI can support language generation, knowledge retrieval, visual and audio analysis, creative tasks, and data processing. Tools range from predictable, rules-based automation to generative AI, whose outputs may vary, lack transparency, omit information, or sound plausible while being wrong.<br /><br />AI does not change who is accountable: auditors retain responsibility for professional judgment, audit quality, and compliance with applicable standards. Firms should manage technology-related quality risks, while engagement teams must select suitable tools and assess their outputs. Risk depends on both the tool and the use case; higher-consequence applications require stronger validation.<br /><br />The course presents a five-step approach: define the procedure’s objective, select an appropriate tool, identify risks, perform procedures to address those risks, and determine documentation needs. Auditors should assess whether outputs are relevant, accurate, complete, and logically appropriate. Language-based outputs may require tracing facts to source materials and considering omissions; data-based outputs may require checking logic, completeness, formulas, and calculations.<br /><br />AI use alone does not automatically require documentation. However, when AI meaningfully assists in performing an audit procedure, the file should explain the tool and objective, inputs and their sources, what the tool did, the output relied upon, how the auditor evaluated it, and the conclusions reached. Research, administrative tasks, and drafting documentation of procedures already performed generally do not require separate AI documentation.<br /><br />The central message is to use AI cautiously, apply professional skepticism, verify outputs, and document AI-assisted audit procedures sufficiently for an experienced auditor to understand the work and conclusions.

Keywords

artificial intelligence

audit work

documentation requirements

professional judgment

audit quality

generative AI

quality risks

output validation

professional skepticism

audit procedures