Ukrainian YouTuber found her AI clone selling Russian goods on Chinese internet

Some of the videos using her likeness were made with AI tools provided by a US-based company launched in China.

Olga Loiek is trying to carve out a life of her own. The 20-year-old Ukrainian studies cognitive science at the University of Pennsylvania. Last year, she landed an internship that has taken her to Germany. She also started a YouTube channel a couple of months ago, where she talks about mental health and shares life philosophies.

Lately, she has become quite a recognizable figure on Chinese internet. But there, she’s not Olga Loiek, a young woman from Ukraine. Instead, she’s a Russian woman who speaks Mandarin, loves China, and wants to marry a Chinese man. Her name is Natasha, or Anna, or Grace, or else.

Loiek was not the only woman who got her image stolen, reproduced and repurposed to be posted on Chinese internet. Around China’s Spring Festival this year, countless videos of foreign looking women made with artificial intelligence have popped up on Chinese social media platforms including Douyin and Xiaohongshu.

In these videos, the AI generated women would almost always identify as Russian and make pro-China and Pro-Russia comments. They also encourage viewers to visit their online stores to shop “authentic” Russian goods.

Some of the videos featuring Loiek’s image seem to have been made on HeyGen, a Los Angeles based AI service that offers tools to create digital avatars for business use. The company was launched in China in 2020 and has since moved to the US.

In recent years, Chinese tech companies have invested heavily in digital avatar and other AI powered technologies. At the same time, concerns about Beijing utilizing AI for its disinformation operations have been growing. A policy expert I talked to for a VOA Chinese story said that it’s only a matter of time before China and Chinese tech companies catch up with the West on AI capabilities.

“This feels very violating.”

Not long after Loiek launched her YouTube channel in late November, she started receiving messages from people telling her that they have seen her speaking Mandarin on Chinese social media. At first, she didn’t think it was all that serious. But seeing the videos herself changed that.

“I started translating the videos with Google Translate and I realized that most of these accounts are talking about things like China, Russia, how good the relationship between China and Russia is,” she told me. “This feels very violating.”

Loiek left her home in Ukraine to study in the US in the summer of 2021. Six months later, Russia invaded her homeland.

“My family lives in the West of Ukraine, which is a little bit more stable at the moment,” she said. “But every single day there's some sort of attacks happening. So, it's pretty difficult.”

Most of the AI videos featuring Loiek’s face look quite realistic, although the awkward body movement, facial expressions and glitches could still give them away. These avatars all seem to be based on how Loiek looked in her YouTube videos. Sometimes even the background looks exactly the same.

In some of the videos, the avatars would talk about how much they value Russia and China’s close ties. “China and Russia are brothers in trouble. Even if the whole world abandons us, China wouldn’t. Long live the China-Russia friendship,” the avatar says in one video.

In other videos, the avatars either praise Chinese history and culture or talk about how much Russian women want to marry to China.

“If you marry Russian women, we will wash clothes, cook, wash dishes for you every day,” an avatar says. “We will also give you foreign babies, as many as you want.”

Douyin, China’s version of TikTok, has labeled some of these videos potentially AI generated. But comments show that many believed they were looking at a real woman. One netizen wrote “Russian beauty, Chinese people welcome you.”

“Seeing my face saying all these things, it was very shocking because I would never, ever in my life say it, obviously, given that I'm from Ukraine,” Loiek said. “This is probably used to make people, maybe people in China, to feel that foreigners feel that their country is superior.”

Loiek said she found at least 32 accounts imitating her on Chinese social media. I found a couple on Douyin as well. Most of these accounts would ask viewers to visit their online stores and shop what they say are authentic Russian goods.

Pretending to be Russian as a business strategy isn’t new. After Moscow invaded its eastern neighbor in 2022, Many Chinese netizens strongly supported Putin and his war. Some influencers started putting filters on their faces to make themselves look more exotic and claimed to be Russian. Many of them harnessed the traffic garnered to bolster their online businesses.

Some videos were made with tools provided by a US-based AI startup

On Bilibili, China’s biggest video site, some of the AI videos using Loiek’s face are marked with the logo of HeyGen, an indication that the video was generated on the company’s website.

In one tutorial posted on Bilibili, the demonstrator even shows how to make a short video on HeyGen with a clip of Loiek talking.

HeyGen is an AI company headquartered in Los Angeles, specializing in realistic digital avatar and voice generating and video translating. The technology developed by HeyGen was used in AI videos of Donald Trump and Taylor Swift speaking perfect Mandarin that went viral on Chinese social media in October.

HeyGen was launched in 2020 by Joshua Xu and Wayne Liang in Shenzhen, China. Xu and Liang graduated from Shanghai Tongji University and later went on to work for social media companies Snap and ByteDance respectively before starting HeyGen together, which was called Surreal at the time.

Three months after launching, Surreal was able to secure funding from leading Chinese investment firms such as Sequoia China and ZhenFund. In November last year, HeyGen reportedly has applied for the deregistration of its Chinese entity. Forbes reported that the company is now valued at 75 million dollars.

HeyGen’s moderation policy states that users cannot generate avatars that “represent real individuals, including celebrities or public figures, without their explicit consent”. The company’s official tutorial video on avatar making also shows that users have to submit a video of people giving consent to the use of their likeness. It’s unclear how some in China were able to circumvent the requirement to make videos of Loiek.

HeyGen has not responded to my request for comment.

Loiek said she and her YouTube subscribers have sent complaints to Chinese social media companies. She said since then 10 of the 32 accounts imitating her have been taken down. On Douyin, accounts that I found have been either suspended or banned from posting. But AI videos of Loiek can still be found on the platform.

Douyin’s parent company ByteDance has not responded to my request for comment.

The Chinese government rolled out provisions to regulate deepfake and other “deep synthesis services” in early 2023. The law prohibits generating deepfakes without the consent of the people whose image or other information is used.

Loiek has posted her story on YouTube. It was shared on Chinese social media as well. Netizens across platforms have expressed sympathy to her and called for tougher regulations on AI.

Loiek said she does not plan to quit YouTube or stop posting.

“Just because there is a threat does not mean that you're not supposed to pursue your interest,” she told me. “But at the same time, we need some sort of regulatory frameworks, so we can understand and we can prevent these things from happening.”

Beijing’s AI Disinformation threat

The introduction of Open AI’s ChatGPT in 2022 has sparked a wave of enthusiasm in artificial intelligence all over the world. Chinese tech giants like Baidu and Tencent have heavily invested in the technology. In China, one of the more hyped-up services powered by AI has been digital humans.

Tencent and Xiaoice, a Chinese AI studio spun off from Microsoft, have offered digital human services that can clone people and turn them into AI avatars at a relatively low price. Digital humans have been favored by many live streamers in China. Using their own avatars to perform for them, the live streamers no longer have to worry about labor cost and work hour restrictions.

But AI avatars have also been found in online disinformation campaigns that spread pro-China and anti-US narratives. In February 2023, research firm Graphika found a social media campaign that promoted the interest of Beijing using realistic-looking computer generated people in videos.

In September, the United States state department in a report also warned: “Access to global data combined with the latest developments in artificial intelligence technology would enable the PRC to surgically target foreign audiences and thereby perhaps influence economic and security decisions in its favor.”

Nathan Beauchamp-Mustafaga, a policy researcher at the RAND Corporation, told me that he believes it’s only a matter of time before Beijing and Chinese tech companies finally catch up with the West.

“It might be a year to two years, but we should assume that they‘ll catch up,” he said. “And then the Chinese government will at some point have some kind of capability similar to open AI.”