On Autocrats, Algorithms, and the Blurring Line

Note:

Written in spring 2024 — my final quarter of graduate school — for GPPS 481: The Political Economy of Authoritarian Regimes. Two response papers, submitted a month apart, building on each other. The first argues that classical theories of information control and selective repression hold up surprisingly well in the digital age — China's Social Credit System is Dimitrov's Bulgaria with better infrastructure. The second goes further: that the line between autocracy and democracy has blurred enough that authoritarian regimes no longer need to pretend particularly hard, because democratic states are doing a convincing impression of them anyway. The Kern and Hainmueller finding — that West German television actually stabilized the East German regime by making life bearable enough not to revolt — is the most unsettling thing I read that semester. I think about it more than I expected to.

There is a particular kind of intellectual discomfort that comes from reading political theory that keeps proving itself right in real time. GPPS 481 — The Political Economy of Authoritarian Regimes — was that kind of course, and spring 2024 was not an ideal moment to be studying how authoritarian governments sustain themselves through information manipulation. The news cycle had opinions about that.

The two response papers I wrote for this course built on each other across the semester, and together trace an argument I found genuinely unsettling by the end: that the tools of authoritarian control are not exotic or alien, but incremental, rational, and increasingly indistinguishable from the tools that democratic governments use.

The first paper engaged with a straightforward premise — that information quality and repression are inversely related. The better a regime's information, the more selective and targeted its repression can be; the more informationally impoverished it is, the more it resorts to blanketed violence out of sheer uncertainty. This is counterintuitive but holds up. China's Social Credit System is essentially Dimitrov's Bulgaria with better infrastructure: the social contract between citizen and state — political quiescence in exchange for economic stability and social order — functions the same way, just mediated through algorithms and app scores instead of informants and block committees.

The second paper went somewhere darker. Guriev and Treisman's argument is that the modern autocrat doesn't need overt violence or ideological extremism to survive — they need to make people believe, rationally but incorrectly, that the regime is competent and public-spirited. They create a façade of democratic governance: elections, economic rhetoric, managed media. The counterintuitive finding from Kern and Hainmueller — that West German television actually stabilized the East German regime by making life bearable enough that people didn't revolt — is the most quietly devastating thing I read that semester. Entertainment as pacification. The enemy of dissent is not repression but comfort.

What the second paper grapples with, and doesn't fully resolve, is the increasingly blurred line between autocratic and democratic governance. Democratic backsliding — in the United States, the United Kingdom, Brazil, South Korea — has normalized surveillance, censorship, and information manipulation in states that ostensibly don't need those tools. Western companies have supplied internet-blocking technology to Belarus and surveillance systems to Libya and Egypt. The playbook isn't secret. And as democracies adopt elements of the authoritarian toolkit, they inadvertently legitimize it, making it harder for citizens anywhere to identify what they're living under.

I submitted the second paper in May 2024. The conclusions feel more urgent now, not less. The most useful thing the course gave me was a framework for recognizing that the sophistication of modern authoritarian control lies precisely in its mundanity — in how little it requires people to notice what's happening to them. Which is, of course, the point.