With the rapid development of AI, deepfake face-changing technology is becoming more and more mature . In China, "Ai face-changing" has caused all kinds of chaos , and some people even use this technology to impersonate Chinese officials to commit fraud! What's the matter? Today, let’s talk about “China’s AI face-changing phenomenon”! Hiho~ Hello everyone, I am Zhiqi! Do you remember "Ma Yilong"? At the end of 2021, this Internet celebrity who looks very much like Tesla boss Musk became famous on Douyin, and instantly became a meme topic. It even became so popular that Musk responded on the social community. The content of Musk's response was also very interesting . Said that he is willing to meet Ma Yilong on the premise that Ma Yilong must be a "real person". After all, Deepfake technology is too advanced now. Although it is still impossible to confirm whether Ma Yilong is a real person , many Chinese netizens doubt Ma Yilong. Yilong is a product of technology because the technology of using AI to change faces and reshape faces is very developed and very common in China . There are even many Chinese Internet celebrities who have become famous by relying on this technology . For example, the Chinese live broadcaster "Xia Ziwei" relies on this technology. She became famous for her facial features that resemble Ruby Lin and her hilarious and accurate imitation of Ruby Lin's classic character "Ziwei". However, she was once discovered by netizens because she forgot to turn on the filter. "In addition, there are a lot of "fake" chaos , such as fake celebrities live broadcasting goods and pretending to be Russian soldiers to earn traffic , and there are even people pretending to be Chinese officials to defraud people . Today we will take a look at the AI face-changing technology in China. What chaos has spawned? So why is AI face-changing so booming in China? How did Chinese officials respond? But before we start today’s discussion, let’s go into some business service time! [Chaos 1: Fake celebrities live streaming to sell goods] In recent years, many "big stars" have appeared on Chinese platforms such as Douyin, Kuaishou, etc. to sell goods live . But many times if you take a closer look, you will find that these big stars are not themselves at all. "Fake celebrities" generated using AI face-changing technology. Of course, there are indeed situations where "real celebrities" go to live broadcasts to bring goods, so many netizens will initially think they are "real people" or at least look like celebrity faces. But the more they look at them, the more It was only after the multi-screen flaw was discovered that it was AI face-changing. Some analysts pointed out that in the fierce competition in China's live broadcast industry, in order to grab traffic , there have long been live broadcast hosts who rely on makeup and performances to imitate celebrities , such as Jay Chou, JJ Lin and other celebrities. It has been imitated , but now it is easier to pretend to be a star. With AI technology, you can directly put on the face of a star. Therefore, the phenomenon of "fake celebrity live broadcasts" has become more common. Some netizens even said that they had been in different live broadcasts. Seeing the situation of several popular Chinese celebrities such as "Yang Mi" and "Reba" at the same time also made some Chinese netizens joke that "ordinary people are no longer worthy of live broadcasts." However, this face-changing technology is not only used in There are also some people who use AI to "change their race" on "Change Stars " [Chaos 2: Fake Foreign Internet Celebrities] Another phenomenon caused by AI face-changing technology in China in recent years is called "Fake Foreign Internet Celebrities" . It is to turn into a white man to attract attention. Especially after the start of the Russia-Ukraine war, due to the widespread support of Chinese public opinion for Russia, pretending to be "Russian" seems to have become the "new traffic password" on China's Douyin. For example, last year there was a Douyin celebrity "Russian Nana" became famous by sharing her life in China and praising China, attracting two million fans and selling a lot of Russian products . As a result, it was later discovered that she was a Chinese "face-swap" The fake Douyin account was once blocked because of this . This year, another Douyin celebrity "Paul Kochaty" claims to be a "Chechen special forces" who can speak Chinese and is fighting on the front line in Ukraine. He will be in the video Sharing his combat deeds , such as seizing the Ukrainian president's car and arresting active US troops, etc., he attracted nearly 400,000 fans in just a few months. The vodka , honey and other Russian products he sold were also sold out . As a result, netizens continued to Various suspicious points were discovered. For example, he spoke with a Chinese dialect accent. The IP address of his account was also shown in Henan Province, China. Even the building he claimed to be the "Ukrainian Nuclear Power Plant" in his video was actually a power plant in Luoyang, Henan. Later, this incident also Chinese media reported that he used an AI tool to change his appearance to white. Shortly after, Douyin blocked the account of "Kochatie". The method of making money through AI face-changing is not just fake celebrities and foreigners. There are even people using this technology to commit fraud , and the pretended targets range from relatives and friends of the victim to government officials [Chaos 3: Fake relatives, fake officials fraud cases] According to Chinese media reports in recent months, China’s “Ai Face-swapping fraud cases are popping up one after another. In May, a technology company boss in China was defrauded of 4.3 million yuan in just 10 minutes. What happened was that the boss received a call from a friend saying he wanted to hire someone. He helped, first remitting a sum of money to the designated account . Because it was a video call with the friend's face and voice , the boss didn't doubt it and he remitted the money . But later he found out that it was not the friend himself but the friend. After the incident in which gangsters stole his WeChat account and used AI to generate voices and looks to impersonate him was exposed , the keyword "AI fraud is breaking out across the country" was once on the Weibo hot search. Later, many similar cases were reported and arrests were made. There is a group of criminal groups that specialize in this kind of AI fraud. There have even been cases of "pretending to be government officials" recently. In July, there was a case in a northern province of China where gangsters pretended to be government officials and made video calls to the victims . Successfully defrauded millions of yuan. According to Chinese media reports, the reason why this type of fraud cases continue to increase is first of all because the "cost of counterfeiting" is very low. The generated materials fed to AI are very easy to obtain , such as cases of pretending to be relatives and friends, as long as through harassment You can get the voice data and photos of the person you are pretending to be by phone , but you can just go to the social platform to get them. It is even easier to pretend to be an official because some officials often appear on TV and you can easily get rich materials . In addition, technically, this kind of The execution threshold of the case is not high because AI face-changing technology has become very popular in China . Relevant teaching and resources can be found everywhere on the Internet [The Rise of AI Face-changing Technology in China] Deepfake technology has been very popular since 2017. China is no exception and is quickly taking the world by storm. In recent years, with the rapid advancement of AI technology, face-changing has become increasingly easier. As early as 2019, a face-changing app "ZAO" appeared in China . Just upload a face-changing app. By taking a face photo, you can transplant your own face onto the characters of various film and television works to create short films . This APP also quickly became popular on Chinese social platforms. It once topped the "free download list" of Apple Appstore in China. Although ZAO later The fever quickly subsided because the "privacy clause" caused doubts from the outside world . However, AI face-changing technology continues to develop. Now, major companies including Tencent and others have launched similar "face fusion" tools . In addition, there are already tools that allow you to "One-click face-changing" like applying filters during live broadcasts or videos . At the same time, various "face-changing" tool kits are also sold on the Chinese Internet. According to Chinese media reports, most of these toolkits use existing foreign open source software and materials and then convert them into Chinese versions, add Chinese materials, etc. Since there is almost no cost, the prices of such tools are becoming more and more "people-friendly". This year A 79-yuan face-changing toolkit has appeared on China Internet , which contains 30 faces and allows you to "transform" directly during video chat. As the technology is easy to obtain and cheap, AI face-changing is becoming more and more popular. Related crimes such as fraud have become popular and have also increased . Of course, this phenomenon does not only occur in China. Similar problems have occurred in almost all parts of the world. Therefore, governments around the world are trying to regulate this emerging technology. Then, Chinese officials How to deal with it? [Chinese Government’s Reaction] According to reports from Western media such as the Wall Street Journal, China is considered a “very advanced” country in legislating to regulate AI. It has always been a big question for global regulatory agencies on how to regulate technologies like Deepfake. Challenges and China has already introduced relevant regulations. At the end of last year, China issued the "Regulations on the Management of Deep Synthesis of Internet Information Services", which explicitly prohibits any person or organization from using deep synthesis technology to produce and disseminate false information and other illegal content. It also requires the provision of such services. Manufacturers must clearly mark the content as "AI-synthesized" and if they find that someone is using their services to produce and disseminate illegal content, they must "take timely refutation measures." However, the Wall Street Journal pointed out that the "illegal content" defined by this regulation is quite broad. It not only includes false information and infringing content , but also emphasizes that the content generated by deep synthesis technology cannot be used to engage in illegal activities such as "endangering national security and interests", "damaging the national image, disrupting social order", etc. Therefore, some Western analysis I believe that China’s actions are so fast mainly to maintain stability. In the past few months, when relevant Chinese authorities were strengthening the review and rectification of “false information generated by AI”, many banned contents were suspected of alluding to the “Tangshan Beating Case” and so on. and other social events that have aroused heated public opinion in recent years . In addition, these analyzes also mentioned that because the Chinese police are currently focusing their main energy on political censorship of AI-generated content and have relatively few resources to deal with AI fraud issues , the Chinese government may not want to Public opinion is too focused on "AI fraud." For example, a while ago, the topic "#AIfraud is breaking out across the country" hit Weibo hot searches and was banned soon after. However, China's official media "Xinhuanet" stated that this topic is "False news" because according to the Chinese police, the proportion of AI fraud cases is very low and there is no so-called "outbreak across the country" [Our point of view] We have made many videos before that have discussed the problems caused by AI counterfeiting technology Including "AI voice fraud" in North America visiting victims of Taiwan's "Deepfake face-changing incident", etc. A while ago, there was also an incident in Taiwan where a fraud group used AI to change faces into Zhou Tanghao and defrauded fans of 500,000. Obviously, this problem is not only It happened in China. With the advancement and popularization of AI technology, there are chaos and criminal problems caused by various AI counterfeiting technologies all over the world . However, we feel that China’s situation is worth studying and observing . This aspect is because their The population base is large , so various usage patterns should change faster. On the other hand, let’s put aside the discussion of “political censorship, stability maintenance” and so on. In terms of the “management methods” provided by China’s new regulations, we think that if To prevent the crime of "AI face-changing", some of their methods are indeed worthy of reference. For example, they did not ban or significantly restrict the development and application of this technology , but established an accountability mechanism for "AI technology service providers" and required them to Adding "watermarks" and other marks on this type of content allows others to know that the content is generated by AI. We feel that, after all, many things are open source and the prohibition or supervision technology itself is indeed difficult . And getting to the bottom of it, the problem is also It’s not about the technology , but about the users and usage methods, so it should be a more practical approach for them to target technology suppliers and generated content . However, this is just our unprofessional thinking. AI technology actually needs to be managed. There are more aspects to consider , so in the end, I would like to ask you how do you think we should deal with the crime problem caused by AI counterfeiting technology ? Finally, if you like today's video, please share it to let more people know about "China's AI face-changing phenomenon"! In addition, you can also click here to read the discussion of "AI Voice Fraud" and "Taiwan Deepfake Face Changing Case" . So, today's Zhiqi Qiqi comes to an end here . See you tomorrow night!
