AI has dramatically lowered the barrier to entry for hackers, as demonstrated by recent attacks on the Mexican government (150GB stolen by one person using Claude Code) and 600+ FortiGate firewalls compromised by a low-skill attacker. AI-enabled attacks surged 89% year-over-year per CrowdStrike's 2026 report, with breakout times as fast as 27 seconds. Traditional defenses like signature-based detection and SAST can't keep up with AI-generated novel attack paths. The proposed solution is continuous AI pentesting — running AI agents against your own application on every deploy to find vulnerability chains before attackers do. The post concludes by promoting Aikido Infinite, a continuous AI pentesting product.
Table of contents
Attackers already have their superpowered toysDefenders need their own upgradeSuit up and goSort: