US charges three men over $129m scheme that faked CEOs’ voices with AI
THE SO WHAT
A $129M deepfake CEO fraud isn’t a proof of concept anymore—it’s enterprise-scale social engineering. Treat any high-value payment instruction as compromised by default and move to multi-channel, out-of-band verification for wire approvals this quarter, not next year.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechDOJ Says Deepfaked CEOs and Lawyers Helped Steal $129 Million From Hospitals, Banks, and Others
Deepfaked executives enabling $129 million in fraud — with one actor allegedly laundering $400 million over three years — means voice and video are no longer valid as primary authentication. Any workflow that approves money movement or sensitive access on the basis of a call or video needs out-of-band verification this quarter, not next year.
Deep & Emerging TechAI Backlash Spreads Across Europe With Data Center Protests
Street-level backlash against data centers across Europe turns AI buildout into a political and permitting problem, not just a capex one. If your roadmap assumes European compute, you now need a parallel plan for community engagement, site diversification, and potentially slower capacity ramps.
Deep & Emerging TechBinance says CFO Kaiser Ng is leaving to "pursue a new venture" days after a $6.3B hack, citing a six-month transition plan, and names Justin Yek as CFO
A $6.3B hack followed by a CFO transition — even with a six‑month runway — is a stress test of treasury controls, insurance coverage, and counterparty risk. If you custody or trade at scale on centralized venues, re-check your exposure assumptions and contingency plans this week.
Deep & Emerging TechThe US let Nvidia sell chips to an empty office in China
If H200-class chips are reaching shell entities and firms tied to sanctions, export controls are now a probabilistic, not deterministic, constraint on AI compute flows. Assume that any capability you build on “controlled” hardware may eventually be replicated by well-resourced actors abroad and adjust your defensibility assumptions accordingly.