Recently, public security organs in multiple regions have released multiple cases of using AI tools to spread rumors. On June 20th, the official Weibo account of the Shanghai Public Security Bureau's Urban Rail Transit and Public Transport Brigade, @ Rail Transit Yaoyao, released a police report stating that two marketing personnel were administratively detained by the police for spreading rumors about a knife injury at a subway station in order to gain attention. One of them, using AI software to generate video technology, fabricated false videos of subway murders and posted false information online, causing a negative social impact. The reporter learned from the interview that convenient AI tools greatly reduce the cost of spreading rumors, enhance the magnitude and dissemination of rumors. AI rumor making presents characteristics such as low threshold, mass production, and difficulty in identification, which urgently need to strengthen supervision and cut off the chain of interests behind it. Frequent occurrence of AI rumors. Journalists have found that in recent years, AI rumors have occurred frequently in many parts of the country, with fast dissemination, low barriers, and strong confusion. Last year, in the incident of a lost girl in Shanghai, a gang maliciously fabricated and hyped up a series of rumors such as "the girl's father is a stepfather" and "the girl was taken to Wenzhou" using the methods of "headline party" and "shocking body". The gang used AI tools and other tools to generate rumor content, and published 268 articles within 6 days through a matrix of 114 accounts, with multiple articles having over 1 million clicks. The threshold for AI rumors is very low. A case handler revealed that during an investigation into a rumor case, it was discovered that the source of the rumor came from a village in a central province. After investigation, a large number of villagers in the village have used generative AI to write online articles. Some villagers have learned through word of mouth that by pasting the searched internet hotspot information into an AI model, new articles can be generated; Publish on some Internet platforms to make money through traffic. Many villagers spontaneously engage in this activity The investigators said. Strong confusion is also a major characteristic of AI rumor making. The Cybersecurity Bureau of the Ministry of Public Security recently announced a case. Since December 2023, a message about "hot water gushing out of the underground in Huayi District, Xi'an City" has been circulating frequently on the internet, with rumors such as "the hot water gushing out of the underground was due to an earthquake" and "it was due to the rupture of an underground heat pipeline". After investigation, the related rumors were generated through AI manuscript editing. The reporter found that the above rumors were written in a news like manner, with mixed content such as "according to reports", "relevant departments are conducting in-depth investigations into the cause of the accident and taking measures to repair it", and "reminding citizens to pay attention to safety in their daily lives". Ordinary people find it difficult to distinguish between truth and falsehood. Professor Shenyang from the School of Journalism and Communication at Tsinghua University stated that the intervention of AI has greatly accelerated the speed of information generation and dissemination, while also increasing the difficulty of judging the authenticity and accuracy of information. AI tools have technically lowered the threshold for illegal groups to spread rumors and greatly increased the magnitude and dissemination of rumors, which may promote the "customized" generation, "precise" dissemination, and "intelligent" dissemination of online rumors. A research report released by the New Media Research Center of the School of Journalism and Communication at Tsinghua University in April this year showed that in the past two years, economic and corporate rumors accounted for the highest proportion of AI rumors, reaching 43.71%; In the past year, the growth rate of economic and corporate AI rumors has reached 99.91%, with industries such as food delivery and express delivery being the hardest hit areas for AI rumors. The flow profit-making reporter found that behind a number of AI rumor making events, the motivation of rumormakers was mainly derived from obtaining the hits and reading rewards given to creators by the Internet content platform, as well as channeling the operation of e-commerce platforms. In a case pronounced by the People's Court of Shangyu District, Shaoxing City, Zhejiang Province in 2023, criminals illegally purchased AI video generation software through illegal channels to attract followers and generate videos of popular topics collected online through AI with just one click, and then shared them on multiple popular video platforms for profit. After investigation, it was found that as of the incident, more than 20 false rumor videos, including "Thick smoke billowing at the scene of the Zhejiang Industrial Park, with eyewitnesses claiming explosion!", were released by the criminals, involving multiple provinces and cities such as Zhejiang, Hunan, Shanghai, and Sichuan, with a cumulative reading and viewing volume of over 1.67 million times. A journalist investigation found that publishing content and gaining traffic on self media platforms has become a business. Articles such as "graphic creation, AI automatically writes articles, single number easy Nissan 500+, multi number operation, Xiaobai easy to start", "no need to draft, AI 100 original articles a day" are widely circulated on the Internet. Taking a large Internet content platform as an example, the unit price of 1000 times of basic reading of published content is 0.12 yuan, but by multiplying a series of coefficients, the revenue of 1000 times of reading can be increased by 10 times. This means that if the reading volume of an article reaches 1 million times, the creator can earn more than 1000 yuan. Some people use AI to spread rumors in order to attract sales of e-commerce platform products. In February of this year, the Shanghai public security organs discovered that a short video of an artist's "unfortunate fate and passing away with resentment" appeared on an e-commerce platform, causing a large number of likes and reposts. After investigation, the video content was completely forged. After the video publisher arrived at the scene, he confessed that he operated a local specialty online store on a certain e-commerce platform. Due to poor sales, he created eye-catching fake news to attract traffic to his online store account. He doesn't know how to edit videos and can easily generate text and videos using AI technology. In April of this year, the Secretariat of the Central Cyberspace Administration issued a notice on carrying out the special action of "Clearing and Rectifying the Bottom Line of 'Self media' Blog Traffic", which requires strengthening the annotation and display of information sources. If information is generated using technologies such as AI, it must be clearly labeled as generated using these technologies. If publishing content containing fiction, deduction, etc., it must be clearly labeled with a fiction label. Experts say that as the threshold for using AI decreases, the problem of AI spreading rumors may intensify; It is suggested that the main responsibilities of the Internet content platform and AI content generation platform should be compacted to block the interest chain behind rumors; Strengthen legal research and application, and increase the illegal costs of rumormongers and spreaders. Professor Yu Guoming from the School of Journalism and Communication at Beijing Normal University suggests that AI content generation platforms should further improve the identification mechanism for generated content and content sources, such as attaching non removable digital watermarks and text descriptions. The Internet content platform can access the large model system of the AI content generation platform to identify and compare data in real time. Liu Xiaochun, director of the Internet Rule of Law Research Center of the University of Chinese Academy of Social Sciences, suggested that the Internet content platform should strengthen the research on the intelligent identification mechanism of AI rumor making behavior, further optimize the flow distribution and revenue sharing mechanism involving AI content, and reduce the profit space for those with ulterior motives to produce rumors in batches. Yu Guoming believes that in terms of rumor refutation, AI can be used to deal with AI rumor. "For the content that has been confirmed as rumor, Internet content platforms can use AI technology to intelligently push rumor refutation content or give risk alerts to users who have viewed relevant content, so as to reduce the destructive power of rumor." Shenyang and other industry experts said that there are certain difficulties in defining whether there is subjective malice in AI rumor generation, identifying the losses of the rumored party, etc., and it is easy to have legal application problems in criminal punishment for AI rumor making and other acts, which makes it difficult to implement the strike effect. "Our research found that out of 86 cases related to AI rumors, only 15.66% were punished as criminal penalties," said Shenyang. Shu Shenglai, a lawyer at Beijing Mengzhen Law Firm, believes that in the face of rapid iteration of technologies such as AI rumors and criminal forms, research on laws and regulations should also keep pace with the times. We need to further clarify the criteria for determining the crime of AI spreading rumors, strengthen the connection between executions, and accurately punish AI spreading rumors. We should ensure that punishment is appropriate for the crime and punishment is appropriate for the offense, and create a clean and upright online space for the vast number of private internet users. (Lai Xin She)