A visual analysis of six state-backed disinformation operations on Twitter, including from China and Russia, and how they have evolved over the last decade
Context & Ripple Effects
Alexa Pavliuc's Medium piece is an early-systematizing document: by visually comparing six state-backed Twitter operations side by side, including China's and Russia's, it turns what had been scattered takedown reports into a decadal record of how these campaigns changed tactics. It lands just before the coverage wave that stress-tested its subject — Stanford's finding months later that China lagged Russia and Iran at driving engagement despite robust propaganda output.
First-order effects
- Researchers and platform trust-and-safety teams gain a comparative baseline across six operations, making divergence between China's and Russia's playbooks measurable rather than anecdotal.
- Twitter becomes the documented common battlefield, which raises pressure on the platform to detect and disclose coordinated state behavior rather than treat each operation as isolated spam.
Second-order effects
- China's response path is visible in the follow-on coverage: an outsourced propaganda machine using private contractors feeding Facebook and Twitter, plus an amplification bot network retweeting diplomats and state media tens of thousands of times — volume substituting for the engagement Russia-style tactics achieved.
- Russia's evolution runs toward infrastructure outside Twitter itself, culminating in the Doppelganger project exposed via FBI affidavit and European media leaks — cloned outlets rather than platform-native accounts.
Third-order effects
- If the pattern holds, state influence operations industrialize: from hand-run accounts toward persistent branded campaigns like Dragonbridge, which layers AI-generated content across multiple social networks — shifting the defense problem from single-platform moderation to cross-network attribution.
- Effectiveness, not presence, becomes the differentiator among state operators; the corpus suggests scale alone does not buy persuasion, pushing campaigns toward cultural fluency and authenticity rather than raw posting volume.
The trend: State-backed disinformation is evolving from platform-bound account farms into industrialized, AI-assisted, cross-network campaigns whose success depends on cultural fluency more than volume.