ServiceNow just patched a vulnerability that should terrify every CISO. Not because it was particularly sophisticated,it wasn’t. Not because it exploited some cutting-edge AI weakness,it didn’t. What makes CVE-2025-12420 terrifying is how it reveals a fundamental truth about enterprise AI that the industry refuses to acknowledge: we’re building AI systems on the same broken security foundations that have failed us for decades, except now the consequences are exponentially worse.
