Definition
The process of checking whether an AI-powered application is safe and trustworthy before installing it or granting it new permissions. Because AI apps can gain new capabilities overnight, security teams are moving toward continuous automated vetting rather than a one-time review. A store rating is no longer a security signal.
Why it matters
One-click installs and self-updating skills mean trust has to be verified continuously, or a benign-looking app can turn into an attack vector.