Spamouflage

Spamouflage, Dragonbridge, Spamouflage Dragon, or Storm 1376 is an online propaganda and disinformation operation that uses a network of social media accounts to make posts in favor of the Chinese government and harass dissidents and journalists overseas. Beginning in the early 2020s, Spamouflage accounts also began making posts about American and Taiwanese politics. It is widely believed that the Chinese government is behind the network. Spamouflage has increasingly used generative artificial intelligence for influence operations. The campaign has largely failed to receive views from real users, although the evolution of its tactics has allowed some Spamouflage accounts to see some success in receiving organic engagement.

History

In September 2019, research firm Graphika published a report identifying a network of spam accounts across different social media platforms making posts supporting the Chinese government and attacking its critics, naming the network "Spamouflage Dragon" due to its tactic of mixing in non-political spam content as camouflage. The network initially targeted exiled businessman Guo Wengui in 2018 and gradually also added criticisms of the 2019–2020 Hong Kong protests. The spam network's content did not receive much genuine engagement or views from real users, and the report concluded at the time that the low quality of the operation suggests that "it was not a state-backed operation".

In early 2020, after becoming mostly dormant for a period of time following takedowns by social media platforms that occurred as a result of the 2019 Graphika report, the Spamouflage network reemerged with a focus on praising the Chinese government response to COVID-19 in addition to posting about its previous topics of interest.

In February 2021, a Graphika report indicated that some accounts in the Spamouflage network had begun to see some success in receiving views from real users. Instead of solely using disposable spam accounts, Spamouflage began using accounts with a veneer of plausibility, developing fictional personas and co-opting the identities of stolen accounts. Spamouflage posts began to be amplified through retweets by Chinese government officials, including Zhao Lijian and then-ambassador to Venezuela Li Baorong. Its posts closely followed Chinese government messaging and had a new focus on aggressively criticizing the United States and both of its major political parties.

In October 2022, Google's Mandiant reported that Spamouflage, which it calls Dragonbridge, was spreading propaganda and disinformation targeting the 2022 United States midterm elections. One video discouraged Americans from voting and cast doubt on the American political system.

In April 2023, the United States Department of Justice (DOJ) unsealed an indictment against 34 officers of China's Ministry of Public Security (MPS) accused of running a disinformation campaign targeting US-based Chinese dissidents. Meta Platforms and private researchers cited by CNN believe that they are linked to Spamouflage, although the DOJ did not explicitly refer to Spamouflage by name.

In August 2023, Meta announced a takedown of nearly 9000 accounts and pages associated with Spamouflage; their threat intelligence report also covered the propaganda campaign's activities on YouTube, TikTok, and other social media platforms. They also indicated that Spamouflage had created a fake research paper blaming the United States for COVID-19 and attempted to publicize it via various social media networks.

In April 2024, Institute for Strategic Dialogue researchers stated that some Spamouflage accounts had begun making posts about American politics under false American personas, a strategy that was previously used by Russian disinformation campaigns. Under this new strategy, Spamouflage accounts pretending to be supporters of Donald Trump received engagement from real users, with one post being retweeted by conspiracy theorist Alex Jones. The researchers indicated that they did not find Spamouflage accounts pretending to be Joe Biden supporters, although the existence of such accounts could not be ruled out. In May 2024, OpenAI removed accounts used by Spamouflage in influence operations.

Countries and people affected

Guo Wengui

Criticizing exiled businessman Guo Wengui, who has become a prominent critic of the Chinese government, was the initial focus of the Spamouflage network, and Guo has continued to be one of the targets of the network. The prolonged propaganda campaign featured numerous political cartoons attacking Guo and associates such as Li-Meng Yan and Steve Bannon in multiple languages.

Overseas journalists

A June 2022 report by the Australian Strategic Policy Institute (ASPI) found that several journalists, mostly women of Chinese descent working for overseas media outlets, were subjected to a harassment campaign carried out using Spamouflage accounts in what the ASPI called an instance of "digital transnational repression".

Canada and United States

Prior to the 2020 United States presidential election, Spamouflage accounts made posts criticizing Donald Trump in relation to the COVID-19 pandemic, actions he took against China, and the George Floyd protests. Although the network made a large volume of videos, they were of poor quality and received few authentic views. Many of the videos had broken English, and the intended audience of the videos was unclear.

In February 2021, a Graphika report noted that Spamouflage, which previously focused on criticizing Trump and then-Secretary of State Mike Pompeo, began criticizing Biden after the new president was inaugurated. The report concluded that the focus of the network was to spread the narrative of "China's rise and America's fall" rather than partisan election interference.

According to Mandiant, in April 2021, Spamouflage accounts called for protests in New York City against Li-Meng Yan, Guo Wengui and Steve Bannon for spreading "rumors" about COVID-19 and publicized an address purportedly belonging to Guo. Although Spamouflage accounts claimed, sometimes using manipulated images, that the protests were successful, there is no evidence that protesters were successfully mobilized.

Before the 2022 United States midterm elections, the Spamouflage network made posts casting doubts on the American political system, pointing to examples of political division and violence as evidence of the United States's decline. Videos they posted portrayed American elections and the American government as being ineffective in improving Americans' lives.

In the lead-up to the 2024 United States presidential election, Spamouflage accounts began making posts about divisive American political issues. Accounts pretending to be American Donald Trump supporters spread false conspiracy theories about Joe Biden and amplified Russian disinformation about the Russian invasion of Ukraine. In contrast to previous Spamouflage campaigns, these accounts received engagement from real American users.

Spamouflage has made posts criticizing American foreign policy decisions relating to Taiwan, Ukraine, and trade restrictions on China targeting semiconductors. It has also capitalized on the Israel–Hamas war to portray the United States as a threat to world peace.

In 2023, fabricated videos in which Chinese dissident Liu Xin is depicted via deepfake to be making baseless allegations of legal and ethical violations by Justin Trudeau, Pierre Poilievre and other Canadian politicians were posted by Spamouflage accounts. Global Affairs Canada and Liu both believe that the videos may have sought to discredit Liu and hurt Canadian politicians' perception of him.

A 2022 Mandiant report indicated that Spamouflage accounts had called for protests against proposed rare earth mines in the United States and Canada. The minerals are of strategic importance in electronics manufacturing, and the United States had started trying to increase domestic production to counter China's control over the field.

Taiwan

Leading up to the 2024 Taiwanese presidential election, the Spamouflage campaign used generative artificial intelligence to create memes and videos featuring AI television anchors to attack pro-independence presidential candidate Lai Ching-te with fabricated corruption allegations. Spamouflage also created a fake audio clip of Terry Gou endorsing another candidate.

Effectiveness

Low quality and limited authentic engagement

The network has largely been ineffective at getting views and engagement from real users. One reason identified for the failure of the propaganda campaign, in contrast to more successful campaigns by Russia and other state actors, is its operators' lack of familiarity with the global information environment owing to the closed-off nature of the Chinese internet. While some new technical tools adopted by Spamouflage operators, such as generative artificial intelligence, may increase the operation's efficiency, they will not necessarily make the operation more effective.

An August 2020 Graphika report noted the poor quality of Spamouflage content. Videos that Graphika discovered contained "clumsy" text-to-speech voiceovers, grammatically incorrect English, and poorly-translated Chinese idiomatic expressions like "cast a chestnut in the fire will burn themselves with fire".

Successes

Spamouflage has been able to disrupt discussions on controversial subjects by taking up space in search results with spam posts. Additionally, new tactics attempted by the network, such as a small cluster of Twitter accounts discovered in 2024 that pretended to be American supporters of Donald Trump, which researchers named "MAGAflage", have allowed it to receive some real engagement. Researchers have expressed concerns that Spamouflage may start to gain more traction due to amplification by authentic accounts or a scaling up of the MAGAflage strategy.

Attribution

The network has been attributed to the Chinese government by Twitter and Meta. Global Affairs Canada has also linked the campaign to the Chinese government. Sources have linked Spamouflage to the Chinese Ministry of Public Security's "912 Special Project Working Group", which was the subject of a 2023 indictment by the American Department of Justice.

An August 2023 threat report released by Meta notes that the timing of the network's activities match up with office hours in China, and that the network displays coordinated activity across multiple platforms. Institute for Strategic Dialogue researcher Elise Thomas notes that the operation's lack of innovation despite prolonged ineffectiveness is characteristic of a government campaign. In August 2023, Jack Stubbs of Graphika indicated that the firm had seen open source evidence pointing to the involvement of the Chinese group indicted by the DOJ with limited confidence.

Initial reports of Spamouflage's activities did not always attribute the campaign to the Chinese government. Graphika researcher Ben Nimmo speculated in 2019 that amateurs or a private firm was behind the campaign. In 2020, Graphika stated that it was unable to determine the relationship between Spamouflage and the Chinese government.

In a response to CNN, Chinese government spokesperson Liu Pengyu denied China's involvement in Spamouflage.

See also

• Chinese information operations and information warfare
• State-sponsored Internet propaganda
• Chinese government interference in Canada
• 50 Cent Party