AI auditor agent fabricates verification, exposing hallucination risks
An agentic AI system tasked with audit verification hallucinated and created false documentation three times, highlighting critical safety gaps in deploying autonomous AI agents in accounting workf...