AI-Generated ‘People’ Push Anti-Western and Pro-CCP Narratives on Social Media: Report

Illustration picture of Twitter. (Kacper Pempel/Illustration/Reuters)

A massive pro-China “spamouflage” network on social media is being used to distort international perceptions on important issues, such as U.S. gun laws, COVID-19, and racial discrimination, according to a recent report.

The tactic is using a sprawling network of computer-generated “people” to promote the Chinese Communist Party’s (CCP’s) narratives and discredit the West, especially the United States. Fake users, or bots, are used on Twitter, Facebook, Instagram, and YouTube, according to a report (pdf) by the non-profit Centre for Information Resilience (CIR), published on Aug. 5. CIR is an independent and non-profit organization which aims at countering disinformation.

The report found that users who intend to have an open, honest discussion or seek informed opinions on social media platforms would not think twice when a bot crosses their feed.

Bots are almost indistinguishable from real people. Though they appear to be human, these bots are simply generated by cutting-edge artificial intelligence (AI). The report also notes that some user accounts were previously owned by authentic humans and not created for the network, but they were hijacked by bots to carry out the CCP’s online propaganda campaigns.

CIR found over 350 fake accounts that use memes or long blocks of text, filled with negative sentiment. The goal is to amplify speculation over the U.S. handling of COVID-19 issues and agitate users over discrimination and gun violence in the United States. These posts are in English and Chinese.

Some accounts circulate visual contents like cartoons and videos to achieve the following: to sow doubts about the U.S. government’s claims of the CCP virus, commonly known as the novel coronavirus, originating from a Wuhan lab, to delegitimize the allegations of human rights abuses in China’s far-western Xinjiang; and to spread disinformation about exiled Chinese billionaire Guo Wengui and Hong Kong virologist Yan Limeng.

The bot-made posts echo Beijing’s public figures, including the editor-in-chief of the CCP’s mouthpiece Global Times, and spokespersons of the Chinese Ministry of Foreign Affairs Information Department.

“Our research shows evidence of a deliberate effort to distort international perceptions on significant issues—in this case, in favor of China,” said Benjamin Strick, the author of the report.

CIR did their research by finding patterns in certain hashtags used on social networks. They found clusters of accounts consistently boosting content and hashtags from a core account. CIR wrote that the ratio of comments, retweets, and likes is “very inauthentic.” By tracking down the central poster, researchers can identify more narratives.

However, the group states that they don’t have concrete evidence to show the Chinese regime sponsors the network. Nonetheless, these influence operations have a similar hallmark found in user accounts that social media platforms have previously taken down.

“We urge the platforms mentioned in this report to investigate the network, formally offer attribution, and take it down. It’s important that those responsible for its existence are exposed,” said Ross Burley, CIR’s co-founder and executive director.

Social media companies have taken actions repeatedly to remove similar networks.

Twitter took down over 170,000 accounts in June 2020 after social analytics firm Graphika identified a pro-Beijing propaganda network, dubbed “Spamouflage Dragon.” The network of fake accounts actively spread posts that praise China’s communist regime, attack Hong Kong’s pro-democracy movement, and the United States.