- Статьи
- Internet and technology
- Rising from the Depths: why the pornodipfake industry is coming out of the darknet
Rising from the Depths: why the pornodipfake industry is coming out of the darknet
The pornographic deepfake industry is gradually withdrawing from the darknet, experts have warned about this. Resources with sexualized AI content are increasingly being found in the public domain, as it benefits them to be visible: the darknet restricts the audience, advertising traffic, and monetization methods. For more information about why the porn industry is getting closer to ordinary users and what risks this trend carries, read the Izvestia article.
Why is the pornodipfake industry coming out of the darknet
Today, resources with pornographic deepfakes are an established market, Valery Sidorenko, a member of the Public Council under the Ministry of Finance of Russia, head of the working group on developing an approach to regulating content created using AI, and General director of the R.I.M. — INTERIUM communication group, said in an interview with Izvestia.
"Such resources are created for various reasons," says the expert. "For site owners, this is primarily traffic and monetization — advertising, subscriptions, paid orders. At the user level, deepfakes can become a tool for harassment, revenge, blackmail, or discrediting a particular person.
The industry's entry into the open Internet, according to the editorial interlocutor, is logical: the entry threshold has dropped sharply, since one photo may be enough to generate, and ready-made services allow you to get results even for a person without special technical skills. At the same time, it is beneficial for such sites to be visible: the darknet restricts the audience, advertising traffic and monetization methods.
Therefore, resources with pornodipfakes use the usual web infrastructure, search engine promotion, CDN, hosting and advertising systems. As a result, a person may encounter such content in a completely familiar digital environment, rather than on specialized underground platforms.
What does the analyst say about the situation with pornodipfakes
The trend towards the exit of the porn industry from the darknet is confirmed by analytical data: the authors of a recent study published in an American scientific journal found 88 publicly accessible sites with sexualized AI images created without the consent of the people represented on them. Moreover, 38 of the identified sites were entirely devoted to such content.
"Today, the pornodipfake industry no longer needs the darknet: the anonymity for which they went there is sold on the open network as a regular service: hidden domain registration and secure mail," explains Vladislav Shurkhovetsky, a lawyer at Lawyers on Dubrovka. — At the same time, suppliers of such infrastructure do not check the content of customers in advance.
According to the source, it is more difficult to place a website anonymously in Russia: a hosting provider has the right to serve a client only after identifying him in accordance with Part 5 of art. 10.2-1 of the Information Law No. 149-FZ. Therefore, sites with pornographic content from neural networks are moving abroad.
The creators of resources with 18+ deepfakes are counting on legal uncertainty by making them publicly available, adds Vadim Khakov, a lawyer at the Baros Bar Association. In many countries, intimate deepfakes without consent are not explicitly prohibited, so you can choose a jurisdiction with mild regulation. In addition, the commercial model needs an audience from search results: 82 of the 88 sites (93.2%) identified by researchers were on the first two search pages — the darknet does not provide such coverage, the expert notes.
What makes it difficult to block resources with AI pornography
As the capabilities of tools to combat sexualized content from neural networks grow, new ways to circumvent restrictions are emerging, explains Alexander Vinokurov, technical account manager at R-Vision. This is a constant technological race: some tools are used to identify and block resources, while others are used to hide them from monitoring systems and keep the site accessible.
"For example, a resource can change domain names, IP addresses, and hosting infrastructure, use redirect chains, mirrors, and new top-level domains," the expert explains. — A separate problem is the speed of the appearance of new sites: by the time information about the resource gets into the blocking database and the necessary measures are taken, it may already change infrastructure or appear at a different address. Therefore, blocking only by domain name or IP address is not always effective.
In addition, according to Alexander Vinokurov, AI-based technologies allow owners of deepfake resources to change content and infrastructure faster, as well as adapt to the detection tools used. As a result, blocking inevitably turns out to be reactive: first, the resource or material needs to be detected and qualified, and during this time copies can already disperse further, adds Valery Sidorenko.
What is the danger of increasing the availability of porn videos
The increasing availability of AI pornography, which is increasingly becoming publicly available, carries a number of risks, according to experts interviewed by Izvestia. As Vadim Khakov explains, in order to become the "hero" of an 18+ deepfake, just one photo posted in the public domain is enough, so any user can suffer from such content.
"It is no longer necessary to steal real intimate photos for blackmail: an attacker can use ordinary pictures from a social network or other open sources," confirms Alexander Vurasko, Director of Development at the Solar AURA Monitoring Center for External Digital Threats. — Therefore, the argument "I've never filmed anything like this, so it's not a threat to me" no longer works.
At the same time, as Vadim Khakov points out, the consequences of deepfakes are comparable to the leak of authentic photographs: loss of reputation and work, ruined relationships, psychological trauma and blackmail. Materials with minors are especially dangerous. In addition, deepfakes undermine the credibility of audio and video evidence - more expertise is required, and processes are becoming longer and more expensive.
How to deal with the spread of pornodipfakes
In order to solve the problem of pornographic deepfakes, it is necessary to work not only with individual publications, but also with the entire distribution infrastructure, Alexander Vurasko believes. On the platform side, this is a clear complaint procedure, a quick review of the appeal and the detection of repeat downloads. On the side of search and infrastructure services, they respond to confirmed violations within their capabilities.
"For sites that systematically earn money from such content, it is important not only to restrict access, but also to stop its monetization,— the Izvestia interlocutor believes. — The authors of the study pay attention to the need to work with this infrastructure.
According to Alexander Vurasko, labeling "created by artificial intelligence" alone is not enough. It informs about the origin of the image, but does not confirm the consent of the person and does not eliminate the possibility of harassment. Checking the user's age does not solve the problem of using someone else's face without permission.
Intimate deepfakes require the fastest possible response mechanism: a complaint procedure, prompt restriction of distribution, preservation of digital evidence, and the ability to prosecute those who deliberately create and distribute such content to harm a person, says Valery Sidorenko. In parallel, the sites need their own tools for detecting synthetic images and re-uploading already deleted material. Labeling alone will not solve the problem: anyone who creates a deepfake for blackmail or humiliation will not voluntarily put a warning on it, the Izvestia interlocutor is convinced.
Переведено сервисом «Яндекс Переводчик»
