Mass AI surveillance of digital communications makes it harder, not easier, to catch actual criminals.
April 1, 2026
Original Paper
When Mass Security Screening Backfires: Causal Modeling for Complex Equilibrium Systems
SSRN · 6385478
The Takeaway
Mathematical modeling shows that scanning everyone's private chats for illegal content backfires because the sheer volume of false positives destroys the actual investigative capacity of the police. Meanwhile, the most dangerous offenders easily adopt evasion techniques, meaning the system primarily monitors innocent people while the real targets disappear.
From the abstract
<p>Mass security screening programs — from police polygraph testing to AI surveillance of digital communications — share an identical mathematical structure as signal detection problems under conditions of rarity, persistent uncertainty, and secondary screening harms. Standard policy analysis focuses on a single causal pathway: classification accuracy. But causal modeling reveals three additional pathways through which such programs may affect outcomes: strategic behavior (deterrence, evasion, s