The $25 Million Deepfake: Why Your Video Calls Can No Longer Be Trusted

A detailed post-incident analysis describing how attackers used real-time AI-generated deepfakes of executives to socially engineer a finance manager at Arup into authorizing $25M in fraudulent wire transfers; the report demonstrates that audio/video-based identity verification is no longer reliable, documents the low cost and high ROI of such attacks, lists other similar incidents, and prescribes immediate and long-term mitigations including multi-channel/out-of-band verification, mandatory physical tokens for high-value operations, behavioral anomaly detection, zero-trust re-design, and industry coordination on standards and regulation.