Definition
A cybersecurity approach that continuously scans an organization's technology (including its AI systems) for weaknesses, checks which ones are actually exploitable, and prioritizes fixing the most dangerous ones — rather than doing one-off security reviews. Vendors are now extending this model specifically to AI infrastructure and agents.
Why it matters
As AI systems multiply the number of moving parts an attacker can target, boards should expect security teams to move from periodic audits to always-on exposure monitoring as the new baseline.