来源：法治网微信公众号 01 法治网研究院：针对近来不断被媒体曝出的换脸、拟声等各种AI技术滥用乱象，您认为，当前整治 AI技术滥用的难点在哪里？ 黄尹旭：当前，AI滥用手段的进化速度可能超过检测技术，深度伪造已从简单的面部替换升级到动态 表情、光影细节的模拟，有些经过对抗性训练绕过现有检测工具。检测系统刚识别出像素异常特征，滥 用者就可能添加噪声、调整参数来规避。 滥用内容的隐蔽性越来越强，可能只是在真实视频中替换几秒对话，或在新闻里夹杂小段AI生成文 字，传统批量扫描技术很难识别局部篡改。 还有，各平台的监测标准、处置流程不一，难以形成全链条拦截。 李瑞：我认为主要难点有三个：一是技术迭代迅猛，随着人工智能技术的飞速发展，深度伪造技术的质 量不断提升，已经日益逼近甚至超越真实水平，而传统的识别模型在这种高速变化面前显得力不从心， 难以跟上技术发展的步伐。 二是责任主体呈现碎片化状态，从最初的训练数据收集、模型托管服务、分发平台运营，到最终的终端 用户使用，整个链条极为冗长且复杂。这种多环节、多主体的分散格局，使得取证过程变得异常艰难， 归责问题也变得错综复杂，难以厘清。 三是治理标准存在明显缺位，现行的《 ...

Group 1 - The article highlights the misuse of AI technology for malicious purposes, particularly in generating false information and rumors [1][2] - A case in Hunan involved an individual using AI to create a fabricated video about a tragic incident, which quickly gained significant online traction [1][2] - The phenomenon of AI-generated content being used as a new tool for spreading rumors is becoming increasingly prevalent, with a noted rise in the scale of such activities [2] Group 2 - The article discusses the emergence of a gray industry surrounding AI misuse, where the lowering of technical barriers facilitates the creation of false narratives [2] - It emphasizes the need for social responsibility among content creators, as the pursuit of online traffic often leads to unethical practices [2] - Legal frameworks are in place to address the generation and dissemination of false information through AI, with potential penalties including fines and imprisonment [3] Group 3 - The article stresses the importance of platforms taking responsibility for AI-generated content, including proper identification and regulation to prevent the spread of misinformation [3] - It calls for a dual approach of technological ethics and legal enforcement to combat AI misuse effectively [3] - The need for public awareness regarding the legal implications of generating false information through AI is highlighted, emphasizing that there are boundaries that should not be crossed [3]

Group 1 - The "Clear and Bright: Rectification of AI Technology Abuse" initiative has been launched by the Central Cyberspace Administration of China since April, focusing on issues such as AI deepfakes and misleading AI content [1][2] - Over 3,500 AI products, including mini-programs and applications, have been dealt with, and more than 960,000 illegal information pieces have been cleared [2] - The next phase of the initiative will target seven prominent issues, including AI rumors and vulgar content, aiming to establish a technical monitoring system and a long-term working mechanism [2] Group 2 - Investigations revealed that criminals are using advanced AI techniques to mislead the public and quickly gain followers, later transitioning to selling products or courses [2][3] - The production of AI-generated content is becoming increasingly accessible, with individuals able to create videos easily without the need for traditional editing or on-camera appearances [3] - Some AI-generated video accounts have amassed significant followings, with individual videos achieving millions of views and strong monetization potential [3]

Core Viewpoint - The "Clear and Bright: Special Action to Rectify AI Technology Abuse" initiative, launched in April 2025, aims to address the misuse of AI technologies that infringe on public rights and mislead the public due to the lack of content identification [1] Group 1 - The Central Cyberspace Administration of China is focusing on issues such as AI deepfakes and the absence of content identification, which mislead the public [1] - The first phase of the initiative has led to the disposal of over 3,500 illegal AI products, including mini-programs, applications, and intelligent agents [1] - More than 960,000 pieces of illegal information have been cleared, and over 3,700 accounts have been disposed of, indicating significant progress in the initiative [1]

Core Viewpoint - The article discusses the increasing prevalence and sophistication of telecom fraud in China, highlighting the collaborative efforts of various government and industry stakeholders to combat this issue through technology and public awareness initiatives [1][2][3]. Group 1: Industry Response - The Beijing Municipal Internet Information Office and other agencies launched the "Douyin Anti-Fraud Alliance" to enhance public awareness and participation in anti-fraud activities [1]. - The information and communication industry has introduced key technological tools, such as an anti-fraud electronic identification for apps, which has led to a significant reduction in fraud cases by up to 90% in pilot programs [2]. - Douyin's risk management team employs AI technology to improve fraud detection and prevention, blocking over 80,000 fraudulent accounts daily and intercepting more than 4 million suspicious posts and comments [4]. Group 2: Fraud Trends and Tactics - Telecom fraud is evolving with new tactics, including the use of AI for deep forgery and the creation of fraudulent applications, making traditional detection methods less effective [4]. - Major types of fraud include "order refund," "fake online loans," "fake investments," and impersonation of public officials, with a comprehensive strategy involving law enforcement to combat these issues [2][3]. - The fraud landscape is characterized by cross-platform operations, where criminals exploit information gaps between different platforms to target victims more effectively [3]. Group 3: Public Awareness and Education - Douyin has appointed several public figures as anti-fraud ambassadors to promote awareness and educate users about fraud prevention [5]. - The platform has introduced features like "customer service verification" and "dynamic verification codes" to help users verify information and protect themselves from scams [5]. - Public campaigns are being conducted across various cities to raise awareness about telecom fraud, emphasizing the need for collective efforts from government, industry, and users [5].

Core Viewpoint - The incident involving a woman losing 60,000 yuan due to a new type of telecom fraud highlights the hidden and technical nature of modern scams, emphasizing the need for awareness and preventive measures [2][19]. Group 1: How the Money Disappeared - The victim's loss may have originated from a Trojan program that remotely controlled her device, potentially installed through disguised messages or links [3]. - Criminals used the Trojan to obtain sensitive information such as bank card numbers and identification, allowing them to initiate transfers without the victim's active participation [4]. - Exploitation of "no-password payment" features on certain platforms enabled fraudsters to bypass secondary verification, facilitating unauthorized transactions [5]. Group 2: Warning Signs of Trojans - Users should be vigilant for signs such as sudden device lag or overheating, which may indicate a Trojan running in the background [6]. - Automatic redirection to unfamiliar websites or app download pages can signal malicious activity [7]. - An increase in unsolicited SMS verification codes, especially from banking or payment services, is a red flag [8]. Group 3: Emergency Response and Daily Prevention - Immediate actions include disconnecting from the internet to prevent the Trojan from communicating with external servers [9]. - Users should freeze their accounts by contacting their bank or using official apps [10]. - Reporting the incident and preserving evidence, such as messages and call logs, is crucial for police investigations [11]. Group 4: Daily Protective Measures - Avoid clicking on unknown links, even if they appear to come from acquaintances [12]. - Disable small-amount no-password payment options in payment settings [13]. - Regularly scan devices for viruses using reputable security software, particularly for Android users [14]. - Properly wipe data from old devices to prevent information leakage [15]. Group 5: Principles for Identifying Fraudulent Calls - Verify any calls claiming to be from banks or e-commerce platforms by hanging up and calling back using official numbers [16]. - Never disclose sensitive information such as verification codes or passwords, regardless of the caller's persuasion [17]. - Reject calls from suspicious numbers, especially those with international or virtual prefixes [18].

Domestic News - The manufacturing Purchasing Managers' Index (PMI) for April is reported at 49.0%, a decrease of 1.5 percentage points from the previous month, indicating a decline in manufacturing activity [3] - The new Private Economy Promotion Law will come into effect on May 20, aiming to support the development of the private sector [4] - The total holdings of gold ETFs in the Chinese market have reached a historical high, as reported by the World Gold Council [6] - The People's Bank of China conducted a 12 billion yuan reverse repurchase operation using a fixed quantity and interest rate bidding method [11] International News - Traders are fully pricing in four rate cuts of 25 basis points by the Federal Reserve by the end of 2025 [1] - Global gold demand in Q1 reached the highest level for a first quarter since 2016, according to the World Gold Council [2] - The U.S. economy has contracted, with a reported GDP decline of 0.3% in the first quarter, marking the first economic shrinkage since 2022 [6]