In China, in a month, it will be mandatory to specify that a video was created using artificial intelligence and that it reports false information, so that it is published in a legal manner. If these mentions do not appear, the creator of the deepfake will be considered a criminal in the eyes of the Chinese authorities and therefore treated as such. The South China Morning Post reports from the CAC:
“With the democratization of new technologies and artificial intelligence in the online video and audio industry, we have identified several risks. Content such as deepfakes can disrupt the veracity of the information disseminated as well as the social order in the country. The artificial intelligence used for this purpose could have a negative impact on China’s national security. For this reason, we will not tolerate any discrepancy “ .
Deepfakes are designed using an artificial intelligence technique that consists of superimposing existing images and videos on other images and / or videos to make anyone say what they want. In early October, California also legislated on the issue of banning these misleading videos.
The governor of California ratified, on October 3, 2 laws against political and pornographic deepfakes. In the US state, it is now illegal to publish a manipulated video to discredit a California candidate within 60 days of an election. The second law aims to regulate the use of this technique in pornography.
Facebook and Twitter are trying to find effective ways to fight this scourge. In this regard, Facebook recently announced its teams would work with Microsoft and MIT to better detect deepfakes and eliminate them. As part of this alliance, Facebook promises to create a dataset that could be used to create better detection tools. 10 million dollars (9 million euros) are invested in this defense project. Objective: fight against deepfakes in anticipation of the arrival of the presidential elections of 2020.
All experts expect it: deepfakes may pose a major problem in upcoming US elections. This process of intelligent face switching is actually an image synthesis technique based on artificial intelligence. Remember, in April 2018, a video of Barack Obama was created using artificial intelligence . It should be noted that for now, deepfakes are mainly used to create pornographic content . Only 4% of the videos published are in a political setting, and often humorous.