# Automation and Repression (Acemoglu, Gitmez & Shadmehr, working paper, 2026) Back to [[Users/@thesupermannx/thesupermannx-x-profile|@thesupermannx profile]] - **Source:** [MIT Economics PDF](https://economics.mit.edu/sites/default/files/2026-06/Automation%20and%20Repression.pdf), dated **Jun 5, 2026** - **Authors:** Daron Acemoglu (MIT, 2024 economics Nobel co-laureate), A. Arda Gitmez (Bilkent University and CEPR), Mehdi Shadmehr (UNC Chapel Hill) - **Type:** theory model (task-based automation plus a global-games model of revolt) ## Main results (theory) - **Over-automation:** decentralized markets over-automate. - **The state's choices:** a "capitalist state" picks among regulation, redistribution and repression. Automation and repression are **complements**. - **When repression wins:** unless the threat of revolt is weak or capital is low, the state prefers repression, and the economy converges to repression in the long run. It also underinvests in new tasks. - **Democracy:** starting from democracy, capital accumulation makes a coup attractive to capital owners. - **Surveillance (Remark 8):** AI surveillance could lower the cost of repression. ## The authors' caveats - **"A first step":** the authors call the paper this. - **Evidence needed:** they say systematic empirical evidence is needed. - **Sources quoted for color:** Kai-Fu Lee, Jack Ma, Hanauer ("pitchforks"), Mo Gawdat, Alex Karp and Marx. ## Posts about it - [[Users/@thesupermannx/Post Analyses/Post Analysis - Automation and Repression (2026-09-04)|Sep 4, 2026 post]]: 1.0M impressions - Possibly the earlier [Jun 29, 2026 post](https://x.com/thesupermannx/status/2071548970963513837) ("MIT's Nobel Prize-winning economist published a model…"). It had no link, so the match is unconfirmed ⚠️