The ISACA AAIATM framework highlights that the use of AI in auditing does not relieve the auditor of their responsibility for professional judgment. The highest risk is "Over-reliance" or
"Automation Bias," where the auditor accepts AI-generated conclusions without independent validation. If the AI makes a false conclusion due to a hallucination or biased logic, and the auditor fails to "check under the hood," the entire audit report becomes unreliable.