According to the latest investigative report from CCTV China’s “Financial Investigation” program, a criminal network is using AI to produce explicit content on a massive and alarmingly sophisticated scale. From celebrities to everyday women, anyone can instantly become a victim of face-swapping technology.
Technology so accessible it’s scary
The scariest part of this problem isn’t the complexity of the tech, but how widely Deepfake technology has spread. Creating explicit videos is no longer just for tech experts. Anyone can do it with a single portrait photo and a few yuan. Criminals cleverly package cheap “porn keyword” bundles for about 9.9 yuan, using English or slang to bypass automatic platform filters. With simple copy-paste actions into popular AI tools, face-swapped videos with perfectly matched expressions and lip-syncing can be produced and published within minutes.

This flood is fueled further by “gray” offline software that leaves no trace and can’t be banned, openly sold with lifetime memberships under 300 yuan. There’s even an entire ecosystem of guides teaching everything from registering foreign accounts to streaming for profit, promising illicit earnings of tens of thousands of yuan daily for participants.
A dark empire with massive profits
Behind these fake videos lies a highly professional and closed criminal supply chain. At the top, leaders provide tools, tutorials, and keyword suggestions under the guise of “AI photo editing” services to lure customers. In the middle, custom order services let users pick celebrities or acquaintances, specify actions and dialogue, with a 24-hour delivery promise. At the bottom, distribution networks use short clips to bait users into buying raw material packages that can reach hundreds of gigabytes.
The operation’s sophistication shows in how servers are hosted overseas and payments are made via cryptocurrencies like USDT, making tracking extremely difficult. Investigations reveal that in just 18 months, such a group can handle transactions exceeding 12 million yuan. This figure highlights how greed corrupts technology into a tool that profits from others’ suffering.
Digital scars that never heal
The harm caused by explicit AI Deepfakes goes far beyond economic damage. For celebrities, it’s a direct attack on their reputation and careers they’ve worked hard to build. For ordinary women, the consequences are even harsher. Fake images spread online can lead to social stigma, cyberbullying, depression, job loss, or even extreme actions. Once these materials circulate on the internet, they can never be fully removed, leaving permanent psychological scars on victims and harming the mental health of young people who might accidentally encounter them.
Legal boundaries and protection shields
Legally, producing and distributing AI-generated explicit content is a serious crime. Offenders face prison sentences over three years for spreading obscene materials and heavy civil liabilities for violating image rights and personal honor. In severe cases, they can be prosecuted for defamation. In recent years, many developers of “one-click undress” software and face-swap video traffickers have been convicted and had all illegal profits confiscated.
To stop this crime’s spread, close cooperation is needed from all sides. Tech platforms must act as the first line of defense by tightening censorship and blocking unauthorized face-swapping tools. Authorities need to aggressively dismantle cross-border crime rings and anonymous payment methods. Most importantly, individuals must raise awareness to protect their personal information and avoid posting high-resolution portraits on public platforms. When violated, promptly gathering evidence and seeking legal help is the only way to defend one’s rights.
Artificial intelligence is a gift of knowledge meant to drive society forward, but it must never become a weapon that strips away human dignity. Facing the shadow of Deepfake, upholding ethical and legal boundaries protects not only others but also the future of all of us in the digital age.
Source: Sohu