Beyond Sampling: Digital Forensics in Corporate Internal Audit
Traditional audit samples a fraction of the data. Digital forensics lets auditors examine everything — even deleted and hidden material — and speak with evidence, not assumption.
The structural weakness of internal audit is sampling: only a fraction of transactions, contracts, and messages are ever reviewed, and fraud hides in the blind spots. Forensic-based internal audit breaks this limit by examining the full population of digital data — including deleted and concealed material — and proving findings with evidence rather than inference.
The core is not the recovery tool but evidentiary integrity. Chain of custody must be maintained from acquisition through analysis and reporting, with hashing, lawful procedure, and reproducibility, so findings hold up in discipline, investigation, or court. Technical analysis and procedural legality must be designed together.
AI adds speed. LLMs select and summarize the issues that matter from a vast body of artifacts, guiding the auditor quickly to the core — broad coverage, fast narrowing. Yet the AI judgment is only a starting point; humans confirm with original evidence. ‘AI finds, humans prove’ is enforced most strictly in forensics, because admissibility is the lifeblood of audit.
Park Jae-hyun systematized this approach in a 2023 KCI paper on building a fraud-detection system using digital forensics in internal audit, and productized it as AI Audit Advisor. He serves as an executive director at HM Company and a technical director of the Korea Data Forensics Society.
글쓴이 · 박재현 (Park Jae-hyun)
LLM·AI 기반 내부감사 · 디지털 포렌식 전문가 · Ethic Code Engineer