AI in audit data analysis: Risks, limitations, and best practices

Conclusiones clave

• AI in audit enhances efficiency but requires human oversight to ensure accuracy, reliability, and audit defensibility.
• Generative AI excels at analyzing unstructured data like emails and documents, helping auditors identify risks and patterns faster.
• AI has limitations in structured data analysis, making it unsuitable for precise calculations and repeatable audit procedures.
• Key audit risks of AI include hallucinations, variability, and lack of transparency, which can impact audit quality and trust.
• Best practices for AI in internal audit include using enterprise tools, protecting data security, and validating outputs with professional skepticism.