Tiktok’s Bold Glamor beauty filter has AI experts worried about reality

Bold Glamour, which has been downloaded more than 16 million times since its launch last month, offers one of the most convincing effects anyone with a smartphone can use to create an enhanced version or avatar of themselves in real time, according to independent filter makers. , researchers and creative technologists.

Unlike other photo filters, Bold Glamor doesn’t falter when there’s too much movement or when a hand passes in front of the user’s face. It changes faces in real time and there are no stray false eyelashes. Users scrunch their faces and furrow their brows because they can’t believe what they’re seeing.

TikTok spokeswoman Alexa Youssefian declined to discuss the technology behind the filter, but users and researchers say the filter reads people’s skin tone, perceived gender and hair color, then determines the amount of makeup, lip plumping and skin tightening for a certain For your SoCal application. influencer appearance. And artificial intelligence may be guiding this process.

“It appears that the filter used Kylie Jenner’s face as a model for a machine learning algorithm and then blended her face with mine,” Laura Gouillon, the creator of the social media filter, told The Washington Post.

The effect is probably the teaching itself makeup techniques, he said. “In previous beauty filters, the mask was draped over the user’s face, so if a hand or hair covered the face, the effect failed,” Gouillon said. “This filter probably uses machine learning technology to blend these features into your face.”

According to Hany Farid, a professor of computer science at the University of California, Berkeley, the filter likely uses “generative artificial intelligence,” a technology that studies vast amounts of images and text samples, often scraped from the Internet, to create new images or words. (Farid was a member of TikTok’s content advisory board between 2020 and 2022.)

Luke Hurd, creator of Snapchat and Instagram filters, he tweeted that the Bold Glamor effect uses machine learning.

“They use something called a Generative Adversarial Network,” he said. “They take an image of the user, then compare it to a dataset of other images, then redraw the face pixel by pixel on the output of the camera stream.”

The filter’s accessibility is what sets it apart from previous tools of visual manipulation, said Halsey Burgund, a creative technologist at MIT’s Open Documentary Lab.

How to try the new AI technology everyone’s talking about

“The ability to believably manipulate yourself, and soon others, in videos has become a commodity [through this effect], and that’s the real jump,” he said. “This technology has lived for a while in labs, companies, and more esoteric settings on the Internet, but now anyone can do it for free and see how it works.”

Bold Glamor videos can be posted online without a label indicating that the user has used the hyperrealistic effect, and researchers have warned that it will become increasingly difficult to believe what is real online.

“The Bold Glam effect is a sign that the technology to create deep fakes will soon become mainstream,” said Memo Akten, a professor of computer art and design at the University of California, San Diego. “It doesn’t mean they’re going to release the technology, just a signal that they might release it.”

Deepfakes are videos or other media that use artificial intelligence to show people doing things that didn’t actually happen.

Big Tech has moved cautiously on AI. Then came ChatGPT.

On TikTok, Bold Glamor videos are published with a tag directly on the video for user transparency, but this tag disappears when videos are moved outside of the app.

Company spokesman Youssefian declined to address concerns about the warning label for videos exported to other apps. “TikTok can only talk about users’ experiences on TikTok,” he said.

TikTok’s parent company, ByteDance, has been under investigation for years because the Chinese-owned company’s data collection could pose a security threat.

Recently, the US, Canadian and other governments have ordered federal employees to delete TikTok from government devices, citing concerns about potential surveillance. At least 28 US states have banned the app from government devices due to unspecified risks to sensitive and confidential data. A bipartisan group of senators introduced a bill on Tuesday that would give the US Commerce Department the power to ban TikTok or other apps in foreign countries.

On Wednesday, White House press secretary Karine Jean-Pierre said the Biden administration has “national security concerns” related to ByteDance.

In addition to the technological concerns raised by the filter, Bold Glamor can contribute to feelings of inadequacy and unrealistic expectations among users, especially teenagers and other young people, said Linda Charmaraman, a senior researcher at the Wellesley Centers for Women.

The poreless, aerosol beauty standard she advocates is a far cry from the real world, without thousands of dollars worth of plastic surgery and makeup.

“Beauty filters, used unsupervised and then internalized by younger people, are worrying,” Charmaraman said. “The Bold Glam filter is very realistic and less obvious that it’s a filter. And that’s why it can be harmful to the self-esteem of younger users, especially women.”

People have posted shocked reactions about how “scary good” the Bold Glamor effect is; some said the effect made them look like they should audition for modeling, and others joked that the filter looked so real it should be illegal. “I look like a completely different person,” wrote one TikTok user.

Apps have been enhancing faces and filtering appearances for years, but this one seemed “gamey,” Akten said. The Bold Glamor effect, he said, is “ominous.”

“It’s a step towards the world we see in science fiction movies, where we can no longer tell what reality is,” Akten said. “As a result, we will no longer know what to believe or whom to trust.”

Heather Kelly and Tony Romm contributed to this report.