Chinese accounts blast Trump, with help from AI-generated pictures
Chinese social media accounts are not happy with President Donald Trump.
A network of accounts on multiple platforms has been criticizing Trump and broadcasting more positive images of Democratic presidential candidate Joe Biden, as part of an apparent campaign to rebuke the White House, according to a report published Wednesday by Graphika, a New York-based research firm.
The network, which Graphika describes as “Spamouflage Dragon,” produces short videos on a near-daily basis on topics ranging from the Trump administration’s decision to prohibit the social media company TikTok in the U.S. to the government’s response to the coronavirus pandemic. Whether the network was connected to the Chinese government remains unclear, Graphika said. Details of the campaign emerge after a U.S. intelligence assessment determined that Beijing was working to reduce the president’s reelection chances.
“The network was active and public, but ultimately low-engagement,” the report stated. “It typically worked by using apparently hijacked or otherwise repurposed accounts on YouTube, Facebook and Twitter to post politically charged videos in Chinese and then used clusters of fake accounts to share and comment on the posts, creating the impression of an organic community.”
The Washington Post first reported on Graphika’s findings.
Videos seemed designed to amplify existing tensions within the U.S.
Messaging from one YouTube account called Malcolm Daly, which claimed more than 1,000 followers, relied on a profile photo of a young woman, and published videos with names like “The American government ignores science, Politicians turn into health experts” and “#USA It is really hard for Americans today.” Some of the videos received only a handful of views, if any, while others brought attention to police brutality, anti-racism protests and U.S. Secretary of State Mike Pompeo.
Then, the network would aim to amplify engagement around some of those videos with other accounts that used profile pictures generated by artificial intelligence tools. The amplifier accounts would sometimes post the content in multiple languages, including Turkish and Arabic.
The activity comes after the same group last year hijacked Facebook, Twitter and YouTube to denigrate pro-democracy protesters in Hong Kong, Graphika found. The cross-platform effort combined spam and travel-related content with political postings that applauded law enforcement in Hong Kong.