\n\n
Regulators worldwide are moving from curiosity about AI in audits to concrete expectations. New guidance from Singapore's accounting regulator, the UK Financial Reporting Council and the US PCAOB now requires firms to validate AI tools, document every AI‑generated output, and retain human professional judgement. Audit committees are feeling the pressure too: a Deloitte‑backed report finds 82% of committees see AI oversight as only emerging, while 70% say technical expertise is the top skill they need. Frameworks such as COSO’s 2026 internal‑control guidance flag risks like model drift and opaque reasoning, pushing firms to develop audit‑ready evidence packs that prove AI safeguards are operating. The next wave will likely see more standardized AI‑audit checklists, wider adoption of jurisdiction‑specific guidance, and tighter oversight of model changes to ensure audit integrity. [[1, 2]]