The most imminent and dangerous AI milestone isn't superintelligence — it's fully automated, end-to-end ransomware operations and an ai-breakout self-evolving worm. AI risk researchers in academia and industry are missing the importance of this issue. We may have already crossed (or be about to cross) a threshold where AI enables criminal actors to execute enterprise-wide encryption attacks with near-100% automation. The implications are potentially catastrophic: enterprises would need a perfect security posture both externally and internally, which fundamentally challenges how security organizations are structured today.
