China is leading the way in regulating deepfakes. Since January 10, the “Regulations for the administration of Internet information services in deep synthesis” have been in force, which regulates the artificial generation of real-looking content with real and virtual people. While Taiwan, England, Wales and some US states are already taking action against the use of artificially created porn depicting real people, the European Union is in the process of amending its Digital Services Act. But China’s move is the first law in force to comprehensively regulate the phenomenon of deepfakes and AI-created content.
Providers of services for generating deepfakes must guarantee data security and must not process personal data unlawfully without the consent of the data subject. In addition, not only deepfakes need to be clearly labeled to avoid public confusion or false impersonalization. The government is casting a significantly wider net, likely to future-proof the law.
“Services that provide intelligent dialogue, synthetic human voices, face generation, immersive mimetic scenes, and other features that generate or significantly alter information content should be clearly labeled,” China’s legislation now requires. “Any organization or person is obliged not to use any technical means to delete, manipulate or hide the logo in question.”
China’s advance on deepfakes
The state Internet Information Office, the Ministry of Industry and Information Technology and the Ministry of Public Security, who drafted the law together, justify the regulation with points that also move legislators in other countries: The new possibilities are from “a few unscrupulous people”. been used to disseminate illegal and unwanted information and to “slander and damage the reputation and honor of others” or to defraud.
However, China’s foray into the future also has an impact on industrial policy. On the one hand, the government hopes to promote the development of its own industry through legal clarity. After all, artificial intelligence is one of the technologies in which China’s Communist Party wants to beat rival USA. On the other hand, the government wants to play a greater role in the development of standards, explains the law firm Dezan Shira & Associates. “More importantly, if successful in China, the new guidelines can provide a policy framework that other countries could build on or adopt.”
Japan has always tried everything possible with electronics – and often the impossible. Every Thursday our author Martin Kölling reports on the latest trends from Japan and neighboring countries.
The irony is that what might be seen abroad as protecting citizens and their privacy also has a clear oppressive note in the Chinese context, with its strict censorship and harsh penalties for alleged rebels. The initiators emphasize that the creation of deepfakes must not be used for activities that are prohibited by laws and administrative regulations. A popular accusation used to lock up system deviants is disruption of public order.
The enumeration of the goals then finally makes it clear that civil rights in China are not at the forefront. It aimed to “strengthen the management of Internet information services in deep synthesis, promote the core socialist values, protect national security and social public interests, and uphold the legitimate rights and interests of citizens, legal persons and other organizations.”
Dealing with deepfakes in the EU
China obliges the providers of these services to comply with the regulations. In its Digital Services Act (DSA), the EU proceeds in a graduated manner according to the size of the company. When it comes to deepfakes, platforms like Google are on the hook once recent revisions to the DSA go into effect.
In December, the European Parliament added the following Article 30a: “If a very large online platform determines that content is generated or manipulated image, audio or video content of existing people, objects, places or other entities or events, and falsely appearing to a person as authentic or truthful (deep fakes), the provider must label the content in a way that informs that the content is not authentic and that is clear to the recipient of the services is visible.” The consent of the person depicted in the deepfake is therefore not required, but the identification is.
In addition, the Parliament added a sentence to Article 63, which is mainly intended to increase the transparency of large advertising platforms. “In addition, very large online platforms should flag any known fake video, audio, or other file.” The online industry medium Unite.ai judged that the legislation is “preparing for the growing practice of ‘legitimate deepfakes'”. In this form, the creators have secured permission and rights for, for example, face swapping in promotional or promotional material. In both China and Europe, these should not remain the last legal steps. Because the technology is still in its infancy.
(jl)
