AI isn’t making code safer. It’s expanding the attack surface.

As Eric Fourrier, GitGuardian CEO, notes, 28.65M secrets were exposed in 2025 as AI workflows expanded tokens, APIs, and machine identities.

Risk has shifted from code to credentials. Remediation is now the bottleneck.

�� Why AI security is shifting beyond code → https://thehackernews.com/expert-insights/2026/03/the-real-problem-isnt-that-ai-cant.html