Audit teams are being asked for assurance over AI before the criteria and techniques have settled. This course gives auditors a working method: what to audit against, how to scope, which evidence is sufficient for systems that are probabilistic and change frequently, and how to report in a way that boards and regulators can act on.
You will compare internal audit, second-line review and external assurance; draw criteria from NIST AI RMF, ISO/IEC 42001, the EU AI Act and the IIA's AI Auditing Framework, applied alongside the Global Internal Audit Standards; and distinguish engagement types such as ISAE 3000 and SSAE-based attestation, SOC 2 reports covering AI services, ISO certification audits and statutory bias audits. You will build an AI audit universe from the inventory, write audit programs for predictive, generative and agentic systems, and test governance, data lineage, validation evidence, change management and vendor controls. Technical modules cover reperforming accuracy, robustness and fairness tests, inspecting prompt logs, guardrail configurations and red-team evidence, sampling, and using AI tools within the audit itself. The course closes with findings, ratings, audit committee reporting, independence and competence. Scenarios are hypothetical. The Wirewalk Institute certificate attests to completion and assessment of this course only and is not a professional audit credential.