X said it has uncovered a Chinese bot farm of approximately 200,000 inauthentic accounts involved in influence operations.
The platform’s Safety team said in an Aug. 27 statement that about 200 accounts within the newly identified network had posted material aimed at influencing U.S. debate over artificial intelligence and energy policy. X did not say how the rest of the roughly 200,000 accounts were being used.
Other technology companies and U.S. authorities have separately documented Chinese influence operations involving tens or hundreds of thousands of fake, coordinated, or repurposed accounts, including networks tied to Chinese state actors and law enforcement.
Twitter, which was rebranded to X in July 2023, disclosed a network of a similar scale seven years ago.
In August 2019, Twitter identified 936 accounts originating from China in what it called a “significant state-backed information operation” targeting Hong Kong’s protest movement. Twitter said it had “reliable evidence” that the coordinated operation was state backed and that it had also suspended a larger spam-heavy network of approximately 200,000 accounts identified to be the most active portions of the campaign. The company said it was about to proactively suspend these accounts before they became “substantially active.”
In 2020, Twitter disclosed another China-linked operation consisting of 23,750 core accounts layered with approximately 150,000 accounts designed to amplify their content. The company said the network spread geopolitical narratives favorable to the Chinese Communist Party (CCP) and deceptive narratives about the Hong Kong protests.
Related Stories
Flooding the Platform
In January, then-X head of product Nikita Bier said the Chinese regime was using millions of previously registered accounts to flood the platform during politically sensitive periods.
“The Chinese government floods X search results with porn whenever there is political unrest—to prevent their citizens from finding out real-time information,” Bier wrote on Jan. 30. “This has been a difficult problem to solve but we are aware & working on it.”
When asked how the flooding could continue despite X’s restrictions, Bier replied: “They have a pool of 5–10 million accounts that were registered before I locked down signup.”
The problem has been familiar to Chinese language users for years.
During China’s widespread anti-COVID-19 lockdown protests in November 2022, searches for major Chinese cities on Twitter were flooded with pornography, escort advertisements, gambling posts, and other spam as people sought information about demonstrations.
Chinese novelist Murong Xuecun later described the experience in a 2023 interview with Rest of World: “Chinese-language Twitter feels like a bus full of political dissidents pulled up to a red-light district.”
Ahead of the June 4 Tiananmen Square massacre anniversary this year, Chinese democracy activist Zhou Fengsuo complained that porn bots were becoming “more aggressive and far more numerous,” drowning out discussion and remembrance.
After X disclosed the latest bot farm, Allen Zeng, co-founder and CEO of the U.S.-based Sound of Hope Radio Network, said the flooding remained widespread.
“They are like locusts and appear everywhere!!” Zeng wrote, urging X to take further action against the spam.
Users have also described relying on their own tools to manage the flood.
Luo Xiang, an overseas social media account known for criticizing the Chinese regime and its policies, said in May that he had blocked about 1,000 accounts in two days, bringing his total to more than 24,000.
In a follow-up post, he described workarounds including browser filtering add-ons, keyword filters, manual blocking, and hiding replies.
Troll Farm Case
U.S. prosecutors have separately detailed how Chinese authorities allegedly used fake social media identities to influence discussion abroad.
The Justice Department in April 2023 charged 34 officers of China’s Ministry of Public Security (MPS) in a transnational repression case involving a regime-run “troll farm” operated by the ministry’s “912 Special Project Working Group.”
The Justice Department said the officers worked from an MPS facility in Beijing and created thousands of fake online personas on platforms, including Twitter to target Chinese democracy activists, spread CCP propaganda, and shape public perceptions of the Chinese regime.
Prosecutors said topics directed by MPS headquarters included U.S. domestic and foreign policy, Hong Kong, Xinjiang, Russia’s invasion of Ukraine, unrest following George Floyd’s death, and the COVID-19 pandemic.
Some of the personas purported to be Americans, giving U.S. users the impression that people in the United States supported positions favored by Beijing, according to the Justice Department.
The department said investigators obtained a guide instructing group members on creating and maintaining fake social media accounts, using temporary email addresses, posting official Chinese regime content, and interacting with users while avoiding the appearance that the accounts were “flooding” a platform.
The group also tracked members’ performance and rewarded officers who successfully operated multiple personas without being detected, prosecutors said.
Networks Spread Across Platforms
Meta said in August 2023 it had dismantled what it described as the largest known cross-platform covert influence operation in the world, removing thousands of accounts and pages associated with the China-linked network commonly known as Spamouflage.
The operation targeted more than 50 platforms, including Facebook, Instagram, X, YouTube, TikTok, Reddit, Pinterest, Medium, Blogspot, and Vimeo. Meta said it linked the operation to individuals associated with Chinese law enforcement.
Google has tracked the same broader operation under the name Dragonbridge. By January 2023, Google said it had terminated more than 100,000 Dragonbridge China-linked accounts over the operation’s lifetime and disrupted more than 50,000 instances of activity during 2022 alone.
By June 2024, Google said it had disrupted more than 175,000 instances of Dragonbridge activity over the network’s lifetime.
Microsoft has separately documented Chinese operators adopting American identities and increasingly using artificial intelligence to produce political content.
In April 2024, Microsoft said CCP-affiliated actors were using deceptive social media accounts to impersonate U.S. voters, engage on divisive domestic issues, and test which issues drew the strongest responses from American audiences.
Five months later, Microsoft documented China-linked personas masquerading as both conservative and progressive U.S. voters, directly engaging users, tagging politicians and commentators, and creating online groups around shared political interests.
Graphika, which has tracked Spamouflage since 2019, reported in July that the Chinese state-linked operation was still using fake Facebook and X accounts to circulate manipulated event flyers aimed at disrupting events critical of the CCP in the United States and Europe.
In an April investigation, Graphika said it traced part of the Chinese state linked Spamouflage operation to inauthentic Facebook accounts bearing the “Salva Store” name. Graphika assessed with high confidence that the intermediary behind the account-selling network was based in Bangladesh.
Graphika also identified Spamouflage linked assets on YouTube, Facebook, Instagram, X, Tumblr, and Bluesky in the same investigation.
