In the advertisement, a woman in a white lace dress makes suggestive faces at the camera and then kneels. There is something a bit uncanny about her; a quiver at the side of her temple, a peculiar stillness of her lip.
However, if you saw the video in the wild, you might not know that it is a deepfake fabrication. It would just look like a video, like the opening shots of some cheesy, low-budget Internet porn.
In top right corner, as the video loops, there is a still image of actress Emma Watson, taken when she was a teenager, from a promotional shoot for the Harry Potter movies. It is her face that has been pasted onto the porn performer’s.
Illustration: Mountain People
Suddenly, a woman who has never performed in pornography is featured in it.
The ads, which directed users to an app that uses artificial intelligence (AI) technology to make deepfake videos, were discovered in more than 230 iterations across Facebook and Instagram, an NBC news investigation led by US journalist Kat Tenbarge found.
Most of the ads featured Watson’s image; some others used the face of actress Scarlett Johansson. The same ads appeared in photo editing and gaming apps available in Apple’s App Store. Lest the message be lost on viewers, the ads make explicit that they are intended to help users create nonconsensual porn of any women they like.
“Swap ANY FACE in the video,” the ads read. “Replace face with anyone. Enjoy yourself with AI face swap technology.”
Similar ads for deepfake services appear directly next to explicit videos on PornHub. Although deepfake technology can theoretically be used for any kind of content — anything from joking satire to malicious political disinformation campaigns — overwhelmingly, the technology is used to create nonconsensual porn.
A report from 2019 showed that 96 percent of deepfake material online is pornographic. That figure might well be increasing.
The ads on Meta and Apple platforms appeared as consumer demand for deepfake pornography is exploding. The surge comes on the heels of a controversy that rocked online video game communities in January, when a popular streamer, Brandon Ewing — who calls himself “Atrioc” — displayed deepfake pornography of several popular female streamers in one of his online broadcasts.
He later admitted to having paid for the artificial porn of the women, who were his colleagues and friends, after seeing an ad similar to those that appeared on Meta and Apple platforms.
The women whose images were commandeered for Ewing’s pornography issued angry and hurt responses; Ewing himself apologized.
However, the controversy seems to have only made the streamer’s overwhelmingly young and male follower base more aware of the availability of deepfake content — and eager to use it themselves.
Genevieve Oh, a researcher who studies livestreaming, told NBC that after Ewing’s apology, Web traffic to the top deepfake porn sites exploded.
That rapid increase over the past few weeks has followed a slower, but still alarming, growth of the deepfake revenge porn sector over the past several years.
In 2018, fewer than 2,000 videos had been uploaded to the best-known deepfake streaming site; by last year, that number had ballooned to 13,000, with a monthly view count of 16 million.
As deepfake revenge porn becomes more popular, the barrier to access is quite low: The app that misused Watson’s face in its ads charges just US$8 per week.
The rapid increase in the number and availability of nonconsensual deepfake porn videos raises alarming questions about privacy and consent in the digital future. How will the huge number of women — and the smaller, but significant, number of men — who are impacted by this new AI-enabled revenge porn manage their reputations and lives? As the technology improves, how will viewers know the difference between fact and AI-generated fiction? How can nonconsensual material be removed when the Internet moves so much faster than regulation?
The example of these apps — and of the men, like Ewing and his fans, who use them — also illuminates something older, and more uncomfortable, about the nature of porn: that men often use it as an expression of their contempt for women, and feel that the sexual depiction of women degrades and violates them.
This is, in fact, much of mainstream porn’s appeal, at least according to the sentiments of many men who consume it: that it enables men to imagine themselves in control of women, and of inflicting pain and degradation on them.
Deepfake revenge porn, then, merely fulfills with technology what mainstream porn has offered men in fantasy: the assurance that any woman can be made lesser, degraded and humiliated, through sexual force. The nonconsent is the point; the humiliation is the point; the cruelty is the point.
There is no other way to understand deepfake pornography’s appeal. It is not as if the Internet lacks sexual content depicting real and consenting adults. What these apps offer their users is specifically and explicitly the opportunity to hurt women by forcing them into pornography against their will.
After Ewing exposed his deepfake pornography to his streaming audience in January, one of the women depicted issued her own tearful video, describing how the malice and violation of the deepfake had wounded her.
In response, a man sent her a picture of her own crying face appearing on his tablet . The screen was covered in semen.
For now, the women and others who are targeted by deepfake revenge porn have few avenues of legal recourse. Most US states have laws punishing revenge porn, but only four — California, New York, Georgia and Virginia — ban nonconsensual deepfakes. Companies hosting the apps are often based overseas, mostly beyond the reach of legal enforcement; the company whose app was advertised on Meta appears to be owned by a parent company based in China.
Meanwhile, more men will begin to use the technology against more women.
“I was on fucking Pornhub and there was an ad” for the deepfake site, Ewing said in his apology video, by way of explaining how he discovered the AI revenge porn site. “There’s an ad on every fucking video for this so I know other people must be clicking it.”
Xiaomi Corp founder Lei Jun (雷軍) on May 22 made a high-profile announcement, giving online viewers a sneak peek at the company’s first 3-nanometer mobile processor — the Xring O1 chip — and saying it is a breakthrough in China’s chip design history. Although Xiaomi might be capable of designing chips, it lacks the ability to manufacture them. No matter how beautifully planned the blueprints are, if they cannot be mass-produced, they are nothing more than drawings on paper. The truth is that China’s chipmaking efforts are still heavily reliant on the free world — particularly on Taiwan Semiconductor Manufacturing
Keelung Mayor George Hsieh (謝國樑) of the Chinese Nationalist Party (KMT) on Tuesday last week apologized over allegations that the former director of the city’s Civil Affairs Department had illegally accessed citizens’ data to assist the KMT in its campaign to recall Democratic Progressive Party (DPP) councilors. Given the public discontent with opposition lawmakers’ disruptive behavior in the legislature, passage of unconstitutional legislation and slashing of the central government’s budget, civic groups have launched a massive campaign to recall KMT lawmakers. The KMT has tried to fight back by initiating campaigns to recall DPP lawmakers, but the petition documents they
A recent scandal involving a high-school student from a private school in Taichung has reignited long-standing frustrations with Taiwan’s increasingly complex and high-pressure university admissions system. The student, who had successfully gained admission to several prestigious medical schools, shared their learning portfolio on social media — only for Internet sleuths to quickly uncover a falsified claim of receiving a “Best Debater” award. The fallout was swift and unforgiving. National Yang Ming Chiao Tung University and Taipei Medical University revoked the student’s admission on Wednesday. One day later, Chung Shan Medical University also announced it would cancel the student’s admission. China Medical
Construction of the Ma-anshan Nuclear Power Plant in Pingtung County’s Hengchun Township (恆春) started in 1978. It began commercial operations in 1984. Since then, it has experienced several accidents, radiation pollution and fires. It was finally decommissioned on May 17 after the operating license of its No. 2 reactor expired. However, a proposed referendum to be held on Aug. 23 on restarting the reactor is potentially bringing back those risks. Four reasons are listed for holding the referendum: First, the difficulty of meeting greenhouse gas reduction targets and the inefficiency of new energy sources such as photovoltaic and wind power. Second,
RSS