Sarah found out about the video through a worried text from a friend. There she was—or what looked like her—in explicit sexual scenes she had never participated in, with someone she had never met. The 32-year-old marketing executive had become the latest victim of a disturbing technology trend: AI-generated fake videos, known as “deepfakes.”
“I felt violated in a way I couldn’t physically fight against,” Sarah told this reporter, asking that her real name be protected. “Someone stole my identity to create a sexual fantasy without my permission.”
The Technology Behind the Trauma
These convincing fake videos use artificial intelligence and machine learning to put victims’ faces onto existing explicit content. The process, called “face-swapping,” needs only a few clear pictures of a person- often taken from social media, to create realistic forgeries.
An expert from the University of California, Berkeley, explains that today’s AI systems need minimal input to create convincing fakes. “What once required expensive equipment and technical know-how can now be done with free apps and basic computing power,” he says. “The technology has advanced faster than our ethical and legal systems.”
Creating these fakes typically involves AI analyzing thousands of images to learn facial features and movements. This trained AI then maps these features onto existing videos, creating seamless fake footage that becomes harder and harder to identify as false.
From Celebrity Targets to Everyday Victims
While celebrities were the main targets when this technology first appeared in 2017, it has rapidly evolved into a weapon deployed against ordinary people. Research from cybersecurity firm Deeptrace found over 14,000 fake videos online in 2019, with 96% being pornographic. By 2023, that number had exploded to over 100,000 videos.
“The misuse of this technology has reached crisis levels in certain online communities,” warns Danielle Citron, law professor at the University of Virginia and author of “The Fight for Privacy.” “We’re seeing entire forums dedicated to turning innocent photos of women into explicit content.”
South Korea has become a particular hotspot for this form of digital attack. Korean police reported a 355% increase in deepfake-related crimes between 2020 and 2022, with most victims being women whose faces were inserted into pornographic content. Similarly, India recorded over 1,000 cases in 2022 alone, showing a growing problem across Asia.
Nigeria has seen a troubling spike in AI-generated fake videos since 2021, according to official sources. These videos are increasingly used for political and financial crimes beyond sexual exploitation. The country’s outdated 2015 Cybercrime Act offers little protection while advocacy groups are pushing for updated legislation that specifically targets deepfakes while promoting digital literacy programs in schools and communities.
Laws Struggling to Keep Up
The legal response to these fake videos remains scattered and insufficient. While Virginia became the first U.S. state to specifically make this practice illegal in 2019, only 16 states currently have similar laws. At the federal level, the proposed DEEPFAKES Accountability Act has repeatedly stalled in Congress.
“The law is trying to catch up with technology that changes daily,” explains attorney Carrie Goldberg, whose firm specializes in sexual privacy violations. “Victims often find there’s no clear legal path when they’re targeted.”
The European Union has taken stronger action through its Digital Services Act, which requires platforms to quickly remove deepfake content or face large fines. The UK’s Online Safety Bill, passed in September 2023, also makes creating and sharing deepfake pornography a crime.
How to Protect Yourself
For everyday internet users, protecting against this type of exploitation requires taking action. Tech experts recommend several important steps:
First, regularly check your social media privacy settings, limiting who can see your photos and restricting profile visibility. “The fewer clear pictures of your face publicly available, the less material potential attackers have,” advises a cybersecurity director at the Electronic Frontier Foundation.
Second, use reverse image search tools like Google Images and TinEye to occasionally check if your pictures appear in unexpected places online.
Third, use two-factor authentication on all accounts to prevent unauthorized access to additional photos. “Many creators of these fakes first try to access private collections before using public images,” Galperin notes.
Finally, learn how to report harmful content on major platforms. Meta (owner of Facebook and Instagram), Twitter, and Google have specific reporting channels for deepfake content, though success in getting content removed varies widely.
Looking Forward
As AI capabilities continue to advance, experts predict both the threats and solutions will evolve. Companies like Microsoft and Intel have developed detection systems that analyze subtle inconsistencies humans can’t see.
“We’re in a race between creation and detection technologies,” notes Farid. “The question is whether our social and legal systems can adapt quickly enough to reduce the harm.”
For victims like Sarah, the misuse of this technology has caused psychological damage that technical solutions alone can’t fix. “Having to prove that something didn’t happen—that you didn’t participate in something everyone can see ‘you’ doing—it fundamentally changes how vulnerable you feel moving through the world,” she says.
As society confronts this digital problem, the core challenge goes beyond technology to basic questions about consent, digital identity, and the increasingly blurred line between reality and manipulation in our shared online spaces
