A comprehensive Harvard study has revealed the vast scale and sophisticated nature of China’s covert online operations aimed at controlling public discourse on social media. According to the research, the Chinese state fabricates approximately 448 million social-media comments annually. However, these posts are not primarily intended to directly confront critics or engage in debates. Instead, the strategy focuses on overwhelming conversations with a flood of positive and distracting content, effectively drowning out dissent and steering discussions away from sensitive or controversial topics.
This phenomenon is widely recognized among Chinese internet users, often linked to the so-called “50-Cent Army,” a term originating from early reports that alleged government-paid commenters earned small sums for posting favorable remarks online. Yet, the Harvard study challenges this simplistic notion, showing that most of these comments are not the work of freelancers seeking minor payments. Rather, they are systematically produced or coordinated by government offices and state employees. These actors post in coordinated bursts, especially when a topic has the potential to provoke offline collective action or widespread public unrest.
The key objective of these campaigns is saturation, not persuasion. Researchers Gary King, Jennifer Pan, and Margaret Roberts, who led the study, mapped out how these online efforts unfold. When a sensitive issue surfaces—such as layoffs, protests, or scandals—the state’s online operatives do not engage by attacking critics directly. Instead, they pivot the conversation toward safe, non-controversial themes. These include patriotic anniversaries, tributes to national heroes and martyrs, slogans celebrating progress, and local community initiatives. The data shows sudden spikes in upbeat, positive posts precisely when online discussions have the potential to escalate into collective action offline.
By flooding the digital space with such content, the Chinese state creates a distracting noise that absorbs public attention and reduces the impact of critical voices. This approach functions as a form of propaganda by volume, where the sheer number of posts overwhelms alternative narratives and diminishes their visibility. Unlike traditional censorship that relies on deleting or blocking dissenting content, this method relies heavily on crowding out unfavorable opinions with a deluge of alternative messages.
This pattern of coordinated posting is particularly significant during moments of crisis. Whether in the wake of a disaster, a scandal, or an unexpected policy change, the swiftest way to blunt public anger is to bury it under a flood of positive or irrelevant content. For example, Microsoft’s threat intelligence reports have documented China-linked influence operators using AI-generated memes, fake social media personas, and fabricated video “news” segments to amplify pro-Beijing narratives and sow doubt in foreign audiences. These tactics have been deployed around geopolitical flashpoints and elections, including in Taiwan, Japan, and the United States.
Taiwan’s recent elections provide a clear example of this external dimension of China’s information strategy. Academic and government investigations conducted between 2024 and 2025 uncovered coordinated efforts to spread conspiracy theories, flood Facebook with misleading posts, and create crowdsourced rumor sites that appeared local but echoed Beijing’s official lines. Taiwan’s security agencies later issued warnings about a persistent “troll army” responsible for millions of misleading messages tied to pro-China networks. This operation involved a complex mix of fake accounts, AI-generated content, and amplification by state media.
China’s state media ecosystem plays a crucial role in extending this information flood beyond its borders. Outlets like CGTN Digital produce videos and short clips in multiple languages, distributing them widely on platforms such as YouTube and Facebook. CGTN’s YouTube channel alone boasts around 3.3 to 3.4 million subscribers and billions of views, while its English-language Facebook page already had over 52 million followers as of 2017, indicating a massive and growing global reach. These official accounts can boost coordinated bursts of content, injecting more visibility and legitimacy into the flood of pro-government messaging.
A vivid example of how this strategy operates in practice can be seen in a factory safety controversy. Initially, a local hashtag begins trending with photos and eyewitness accounts highlighting the accident. However, within an hour, the same hashtag is dominated by posts celebrating a patriotic commemoration and a neighborhood volunteer campaign, filled with emojis and positive slogans but no mention of the incident. While the original critical voices do not disappear entirely, they are effectively smothered under the weight of the flood of positive posts. This pattern of volume spikes authored mainly by government-linked accounts is precisely what the Harvard researchers documented.
Importantly, the study clarifies why the conventional narrative of “paid commenters” misses the broader picture