Matthew Butterick argues that AI, by amplifying capital concentration, quietly erodes liberal democracy without needing malicious actors. He frames the risk as a natural consequence of AI’s inherent political nature, echoing Winner’s thesis that technology can embed and reinforce power structures.

Butterick cites generative‑AI data practices that expose authors’ works to training sets, prompting lawsuits that challenge the legality of such data use. He notes that even in the near term, AI’s efficiency gains will deepen existing economic disparities.

The piece warns that once AI becomes entrenched, political debate shifts to inevitability, marginalizing dissent. Without early intervention, the technology may lock in a new political‑economic configuration that sidelines ordinary economic participation.