A “deepfake” video series purporting to track a Russian special operations soldier’s adventures in Ukraine was actually created by a TikTok figure in China. The series was used to promote narratives of Russia-China friendship to hundreds of thousands of followers on the Chinese version of the social media site.
Officials overseeing the Douyin app — the only version of the social media platform permitted in China — have recently cracked down by suspending the account on grounds the AI-created videos were being used to spread misinformation, according to a report by Britain’s The Sunday Times.
Authorities say an unidentified suspect, who calls himself Baoer Kechatie on his Chinese TikTok site, used deepfake technology to dupe his 400,000 followers of the Douyin app inside China. The account was first reported this week by Sixth Tone, a Chinese state-owned online magazine, which said a “content creator” used the increasingly sophisticated video technology and AI tools to manipulate his appearance and claim he was a Russian special ops fighter reporting from Ukraine.
“Some accounts have been posting videos, claiming to be from Russia and soldiers at war. In that time, they spread false information such as ’battlefield videos’ and ’battlefield movements’ to attract attention and gain traffic,” Douyin wrote in a statement on Saturday.
The development has triggered fresh unease over the proliferation of deepfake proliferation on social media.
It also comes amid growing concern in Washington over TikTok, a wholly owned subsidiary of the Chinese technology firm ByteDance, which also owns the Douyin app. Both the Trump and Biden administrations have expressed concern that Chinese intelligence agencies could exploit TikTok to gather intelligence on Americans by forcing the parent company to hand over data on some 150 million U.S. users of the social media platform.
Chinese officials say such fears are unfounded. However, a bipartisan group of U.S. lawmakers have been pushing legislation to try and thwart data theft by foreign adversaries via social media. Montana in May became the first state to ban the use of TikTok within its borders. Experts say the ban may be hard to enforce or defend in court.
News of the deepfake operation on the Chinese version of TikTok also coincides with the circulation of a recent Pentagon Defense Science Board report warning that China and Russia are engaged in new kinds of “gray zone” warfare against the United States.
Gray zone operations are defined as the use of military methods short of open combat and employing an array of tactics, including economic coercion, cyber espionage, disinformation and “unattributed” military forces.
The Sunday Times report said the deepfake videos of a Russian soldier featured a bearded man in military fatigues, who just happened to speak Mandarin, but appeared in videos from locations across Ukraine to issue regular updates from the battlefield.
“Hello, Chinese friends,” the “soldier” tells his Chinese admirers in one video. “I’m from Chechnya, Russia. Behind me is a nuclear power plant in Ukraine. We just conquered this place and caught some big fish. One of them was a U.S. consultant.”
In other videos, the purported Russian soldier claimed to have clashed with U.S. Marines fighting Ukraine — no American Marines have been deployed — and to have captured Ukrainian President Volodymyr Zelenskyy’s car.
The paper reported that the deception was exposed following the growing popularity online for his reports. His internet address turned out to be located in China’s Henan Province. The deepfake videos have since been removed from Douyin and the account where they had appeared on the social media app has changed its name to Wang Kangmei, meaning “Resisting the U.S.”
Chinese investigators say Baoer Kechatie, the website operator, was using the bogus Ukrainian reports to increase web traffic in part as a way to sell merchandise through his site, including imported Russian honey, vodka and powdered milk.
• Guy Taylor can be reached at gtaylor@washingtontimes.com.
Please read our comment policy before commenting.