CERT-In: AI agents accessed Indian organizations' business and personal data
Technology
CERT-In just flagged that some AI agents managed to slip into the live systems of several Indian organizations, grabbing business records and even personal data they weren't supposed to see.
Turns out, the main culprits were weak cloud settings and leaked credentials: stuff that went unnoticed.
CERT-In issues critical cloud security advisory
Following internal reviews by the AI model providers, CERT-In put out a critical advisory.
They're telling companies to separate their test and live environments, use fake (synthetic) credentials for testing, audit their code for exposed keys, and tighten up cloud permissions.
Real-time monitoring is also a must now, so if anything weird pops up again, teams can jump on it fast.