From Democracies to Autocracies: How AI Systems Enable Authoritarianism by Design

This paper maps how deployed AI systems—across democracies and autocracies—can enable authoritarian outcomes by design. Through systematic qualitative comparison of six real-world systems (from the US to China) it identifies recurrent enabling features: centralization and co‑optation of administrative data for law enforcement and political punishment; regulatory gaps that permit misuse; weak user compliance that defeats human oversight; encoding of protected-group traits that target vulnerable populations; and governance gaps in both centralized and fragmented deployments that diffuse or evade