The European Union proposed to mark the content produced by neural networks

The European Commission (EC) proposes to label all content created by neural networks. Politico writes about this, citing sources in the commission.

The EC has already developed a legislative proposal for labelling. According to the source of the publication, the marking is necessary in order to reduce the risk of using deepfake technology for any manipulations. In addition, the commission intends to ask companies to develop technologies that would help determine whether the content creator is a person or a neural network. We are talking about different content formats – video, audio, image or text.

“Now this content is mostly a game, but tomorrow you can see a fake pornographic video with a presidential candidate,” – declared source (quoted by TASS).

First, companies will be asked to voluntarily apply the label. Subsequently, it is proposed to include this provision in a voluntary code of conduct in the digital environment, which was concluded by large Western companies in 2018. After that, labeling may be made mandatory. Information about this will be contained in the Digital Services Act. It is planned that this document will regulate the digital space of the European Union from 2024.

In Russia, discussions are also underway on labeling content created by neural networks. The problem of deepfakes was outlined in October 2021 by Deputy Prime Minister Dmitry Chernyshenko. He emphasized that the use of deepfakes without the consent of the people themselves is unacceptable, and added that technologies that help recognize the substitution exist. Head of the State Duma Committee on Information Policy Anton Gorelkin in January toldthat such a measure could appear before the end of 2023. In June it became known that the government did not support a bill introducing separate criminal liability for the dissemination of deepfakes. The author of the amendments was the deputy from the Liberal Democratic Party Sergei Leonov.