Speaking in slightly accented Mandarin, the woman declared that Chinese men should marry “our Russian women.” In other videos on Chinese short-form platform Douyin, she describes how much she loves Chinese food and peddles salt and soap from her home country. “The Russians will not deceive the Chinese,” she assured.
But her lip movements didn’t quite match the footage, which was recently posted to an account called “Ladina.” That’s because it’s a video by Australian professional strategist Shadé Zahrai, who has more than 1.7 million TikTok followers, that has been modified using artificial intelligence. Some people dubbed Ms. Zakhlai’s video clips in Mandarin, making it appear that she was selling Russian products.
Welcome to a booming genre on Chinese social media: AI-manipulated videos that exploit young, purportedly Russian women to drum up support for Sino-Russian relations, stoke patriotic fervor or make money — sometimes all three Of.
It’s unclear who is behind many of the videos, but most end up directing viewers to links to products, suggesting their primary purpose is commercial. The main target audience appears to be nationalistic Chinese men.
These videos often carry hashtags such as “Russian wives” and “Russian beauties.” The women in the book describe the achievements of Chinese men or plead with them to save them from poverty or a less idyllic country.
Another set of videos featured a blonde woman describing her gratitude for coming to China.
“I really envy my Chinese friends. You were born with the most precious identity and the most profound and fascinating language in the world.
Another video showed a woman thanking the Chinese people for supporting Russia through economic difficulties by buying Russian chocolates from her. Over the past year, the whole world has been boycotting Russia, imposing various restrictions and difficulties on us. China is like a savior,” she said.
The videos look more natural, with the woman’s lips synchronized with fluent Mandarin. But they are also fake. They’re based on YouTube videos posted by college student Olga Loiek, her real-life videos about self-improvement and her gap year in Germany.
Ms. Loick does not speak Chinese. She said in an interview that she would never praise Russia in such a way. She is from Ukraine and some of her relatives are still there.
The makers of these films are trying to capitalize on the market created by China’s current geopolitics, technology and public sentiment.
Relations between Russia and China have deepened significantly in recent years, with leaders Vladimir V. Putin and Xi Jinping declaring an “unrestricted” partnership in the face of growing hostility from the West. Putin visited Beijing last week, where Xi Jinping welcomed him with fanfare.
The use of foreign faces to praise China is also intended to stimulate national pride or nationalistic sentiments among Chinese audiences. In a censorship environment where more and more topics are off-limits, nationalist content has become one of the most reliable drivers of online traffic in China.
Zhang Chenchen, a professor of international relations at Durham University in the UK, said that this kind of nationalism – like nationalism around the world – often contains some kind of gender discrimination.
“This display of sexual objectification of young white women is a classic trope of gender nationalism or nationalist sexism,” Professor Zhang wrote in an email. “Viewers can reaffirm their ethnicity while watching this content ism and male pride.”
In several videos featuring Ms. Zahlai’s likeness, the fake character referred to her viewers as “Big Brother.” The figure also noted that Russia would not sell the products in Japan or South Korea, two countries with tense relations with China.
The Chinese government often encourages online nationalism, but there is no indication that this has anything to do with deepfake videos (although some local governments have teamed up with real Russian women to promote similar messages about China’s appeal). There is also a small group of genuine Russian influencers on Chinese social media, many of whom are young women.
Many video makers may simply be trying to capitalize on China’s enthusiasm for shopping through live streaming and short videos. As AI technology becomes more advanced, some Chinese companies have switched from real to virtual salespeople to save costs.
Lu Haibing, a professor at Santa Clara University who studies artificial intelligence governance, said that because artificial intelligence technology is developing so quickly and becoming more accessible to the public, artificially generated videos may become more common as a sales strategy .
Ms. Zahlai’s management company said in an email that the AI modifications were of “poor quality” and that they “may appear fake even to a casual viewer.” Some of the account’s videos had only a few dozen views, but one discussing marrying a Russian woman had 22,000 views.
This doesn’t seem to matter. An automated counter that popped up in one of the account’s videos indicated that the salt brand being promoted had been purchased 360,000 times on the platform.
When The New York Times contacted the TikTok account with the manipulated videos of Ms. Zahlai, the account holder confirmed in a message that he had produced the videos. “You set three things: audio, video and mouth. You can make any video you want,” he said before unfriending the reporter.
The level of complexity varies. Some of the fake women look entirely computer-generated, with stiff movements that make them look like glorified Sims. Some of the pieces, like those similar to Ms. Loick’s, are very good.
“Even though I knew it wasn’t me, the realism was scary,” Ms. Loick said. “When I decided to create my YouTube channel, I was aware of the dangers of deepfakes, but I believed this was primarily a concern for prominent political and entertainment figures. Now I realize that anyone with a video of themselves online could be affected.
Ms. Loiek reported the accounts on Xiaohongshu and made a YouTube video about her experience. Eventually, most of the profiles using her likeness were shut down.
Social media platforms have stepped up censorship in recent weeks, removing videos about artificial intelligence or labeling some of them. China was the first country to enact regulations for generating artificial intelligence, and on the face of it, some of its policies are stricter than those in the West.
But countries around the world are struggling to enforce their rules. Spotting inappropriate behavior can be particularly difficult in China, where many foreign social media outlets are banned due to the country’s closed online environment.
Foreign influencers are unlikely to know their images have been used on Chinese social media and file copyright complaints. A 35-year-old man said that Chinese platforms may also not exploit overseas content when checking artificial intelligence manipulation. The man, who gave only his surname, Chen, said he earned about $1,000 a month from the accounts before he closed them in March over concerns about increased regulation.
But more are still surging. Professor Lu of Santa Clara University said Russia may be a hot topic now, but the practice could soon spread to the next trend.
“The people behind this will manipulate any topic possible to get people’s attention,” he said. “Showing parents ‘how to get into top schools’; to young ladies ‘how to be beautiful’. I believe that in the future everyone will use AI technology to customize topics so that videos can attract specific audiences.
