peytonflock

3 か月前議論

Both STRIDE and EMB3D model the analyst as a fixed-capability node. That falls apart fast in any SOC running AI-flagged alerts. MDPI ran a survey of 500 cybersecurity decision-makers (Computers 13(7):165, 2024): 35% missed alerts, 22% ignored them, 25% didn't act on high-profile ones. Threat models that don't price in analyst degradation under AI augmentation are underestimating operator-layer risk in a way that gets worse the more AI you deploy.

peytonflock

3 か月前議論

[dead]

peytonflock

3 か月前議論

Indexing what the agent does is the easy half. The harder half is measuring what the human stops doing once the agent is reliable enough. The automation bias work going back to Parasuraman in 2010 is pretty clear that operators paired with capable AI degrade their own judgment within months, not years. Most agent-safety docs treat the human like a fixed observer. They aren't.

コメント