Naked deception: how will Russia deal with AI pornography
In Russia, the first sentences were passed for creating pornographic materials using neural networks. In the Bryansk region, a 17-year-old boy received two years of probation for distributing 18+ dipfakes based on photos of a minor victim. The second person convicted of a similar crime was a 48-year-old man who used images of nine people as references for neural networks, and then distributed "adult" content on social networks. Read more about these sentences and the prospects for combating pornographic deepfakes in Russia in the Izvestia article.
What is known about the first sentences for pornographic deepfakes
The press services of the Bryansk and Lipetsk regions reported on June 4 that the first sentences were passed in Russia for creating pornographic materials using neural networks. In the Bryansk region, a 17-year-old boy was sentenced to two years of probation for distributing porn generated by a neural network.
As the court found, the defendant received from an acquaintance photos and videos of an intimate nature with the participation of a minor victim. After that, he used a neural network to create pornography with their help, and then sent the materials in one of the messengers "as a joke" on the victim.
During the trial, the teenager fully confessed. The court found him guilty under the article on the distribution of materials with a pornographic image of a minor using the Network. In addition, compensation for moral damage in the amount of 50 thousand rubles was collected in favor of the victim.
In the Bryansk region, a 48-year-old man was convicted of a similar crime. He created a social media group with erotic images made by neural networks. However, he used photographs of nine people as references for artificial intelligence. The man fully admitted his guilt and repented of what he had done. The court sentenced him to three years of probation and banned him from administering social media groups for five years.
Why pornographic deepfakes are becoming a problem in Russia
The main problem of pornographic deepfakes is related to the accessibility of this phenomenon, says Alexey Drozd, head of the security department at Serchinform, in an interview with Izvestia. First, modern neural networks are able to work independently, and not just as a service: users can avoid paying for a subscription and generate content safely, thereby maintaining anonymity.
— Secondly, increased productivity while reducing hardware requirements: you can create such materials even on ordinary computers, not just on expensive servers, — says the expert. — This significantly expands the circle of people who are able to produce them.
Finally, according to the source, the quality of the created deepfakes has improved, and the artifacts are minimal, which makes the fakes convincing to ordinary people. In turn, Vladislav Shurkhovetsky, a lawyer at the Lawyers on Dubrovka International Criminal Court, notes that from the point of view of criminal law, there are no problems in the production of such videos in private (for himself), however, a lot of areas of criminally punishable application are assumed related to the further spread of such digital "creativity."
In particular, we may be talking about Articles 128.1 ("Libel"), Article 159 ("Fraud"), Article 163 ("Extortion"), as well as Articles 242 ("Illegal production and trafficking of pornographic materials or objects") and Article 242.1 ("Production and trafficking of materials or objects with pornographic images of minors") of the Criminal Code of the Russian Federation, as in the cases of the Bryansk and Lipetsk regions.
What are the risks for victims of pornographic deepfakes?
The main danger of deepfakes lies in the total compromise of personality, says Maxim Fedosenko, a leading engineer and analyst at the Gazinformservice Cybersecurity Analytical Center. In other words, neural networks create a "digital clone" of sufficiently high quality, which can be difficult to identify unambiguously using standard methods, which causes a big blow to the reputation of the victim.
"If we talk about intruders, then for them deepfake technology has long gone beyond the tool of harassment, being an extremely effective lever for social engineering: a victim who finds himself deceived by neural networks becomes as vulnerable as possible to blackmail, financial fraud and targeted attacks," explains Izvestia's interlocutor.
Current AI-based technologies are capable of producing photographs that cannot be distinguished from the real ones, adds Vladislav Shurkhovetsky. Challenging and proving the faking of such images can result in years of litigation and large expenses for experts with the prospect of a result in the form of compensation for moral damage in the amount of 50 thousand rubles, as in the case from the Bryansk region.
Meanwhile, Vadim Khakov, a lawyer at the St. Petersburg Bar Association, evaluates the first verdicts in cases of pornographic deepfakes cautiously and positively.: The very fact of their issuance signals to society that the state takes the problem seriously. However, the practice is still extremely heterogeneous: courts have difficulty proving intent, identifying the author of the content, and distinguishing between related parties.
"These sentences are only the first in the emerging judicial practice and cause a lot of discussion in the legal community, but the fact remains that the courts have ready—made regulatory tools for qualifying such actions, which positively characterizes the relevance of our criminal legislation and law enforcement agencies," notes Vladislav Shurkhovetsky.
What are the prospects for combating pornographic deepfakes?
Criminal penalties are a necessary but insufficient tool to combat pornographic deepfakes, Vadim Khakov believes. According to the defender, actual imprisonment will certainly have a deterrent effect on some potential criminals, however, criminological practice shows that where latency is high and the probability of being caught is low, the severity of punishment compensates for its inevitability only partially.
"The key shortfall is not in the severity of sanctions, but in disclosure: anonymous accounts, VPNs, and foreign servers make it much more difficult to establish identity," explains Izvestia's source. — At the same time, it is necessary to develop technical measures — mandatory labeling of synthetic content and the responsibility of platforms for its distribution.
According to Alexey Drozd, one should not forget about the "traditional" mechanisms for bringing to justice those who have already committed a violation: when a platform or messenger provides certain data to law enforcement upon request as part of an investigation. But here there is a difficulty in international cooperation. Many sites are outside Russian jurisdiction and may ignore requests from law enforcement agencies.
Currently, the legislature is considering a bill to expand art. 272.1 of the Criminal Code of the Russian Federation ("Illegal use and (or) transfer, collection and (or) storage of computer information containing personal data, as well as the creation and (or) maintenance of information resources intended for its illegal storage and (or) distribution"), this topic is periodically discussed at meetings, Vladislav recalls. Shurkhovetsky.
— However, this will only work in a complex: a separate structure in the Criminal Code of the Russian Federation, which implies severe punishment, the obligation of platforms to remove material on the complaint and their responsibility for non—fulfillment of these actions, simplified recovery of moral damage and adequate, rather than symbolic amounts of compensation and technical ways to recognize and promptly remove forgery, - says the lawyer.
Technical methods seem to be the most effective, as they can be as efficient and devoid of bureaucracy as possible, assuming that a permanent and high-quality AI-based system will do this, concludes Vladislav Shurkhovetsky.
Переведено сервисом «Яндекс Переводчик»
