[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fT5s73xqUVIAaQccum1GMOEvNc7g_T9FNvEkrAntKoBo":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1347,"明起AI生成内容必须“亮明身份”","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">国家网信办等四部门联合发布的《人工智能生成合成内容标识办法》，9月1日起将正式施行，所有AI生成的文字、图片、视频等内容都要“亮明身份”。针对这些内容在传播过程中该由谁来管？怎么管？\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">《人工智能生成合成内容标识办法》明确，人工智能生成合成内容是指利用人工智能技术生成、合成的文本、图片、音频、视频、虚拟场景等信息。人工智能生成合成内容标识包括显式标识和隐式标识。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">平台在服务提供者的内容上架或上线时要进行审核，核验生成合成内容标识，对未标识或疑似生成内容要添加风险提示，从而在传播端阻断虚假信息扩散。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">中国人民大学未来法治研究院执行院长 张吉豫：“标识办法”中进行检测分三个层次，如果平台是从元数据中检测出隐式标签，那么平台就可以标识这个内容是AI生成的，但是如果它只是通过一些算法推测这个内容是AI生成的，它应该是标识为疑似AI生成。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">针对目前由于技术的局限性而导致的部分原创内容被误判为AI生成，创作者权益受损等情况，法学专家指出，当前的算法仍然存在数据偏差和场景适配不足等问题，在误判率难以归零的情况下，应当建立相应的保障机制。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">张吉豫：未来相关平台进行了分级标识后，其实也应该建立分级的处理办法，比如说对于明确它是AI生成的，通过隐式标识确定的，应该采取更严格的措施，但是对于很多应该标识为疑似AI生成的，应该提供更便捷的用户反馈渠道，也不宜对用户进行严肃的惩罚。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">法学专家建议，首先网络服务提供者要公开算法原理及误判样本数据，提升透明度；其次要设立申诉渠道，允许用户提交证据申请复核，从而保证监管与创作自由的平衡。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">让生成内容看得见查得着 以规范促发展\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">据记者了解，截至目前，我国已有490余款大模型在国家网信办完成备案，240余款大模型在省级网信办完成登记，我国生成式人工智能产品的用户规模已达2.3亿人。“标识办法”的施行，会对行业、个人用户，以及相关技术的发展带来哪些影响？\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">据记者了解，作为我国人工智能领域首部“办法+强标”的组合拳，强制性的国家标准《网络安全技术人工智能生成合成内容标识方法》将于9月1日与《人工智能生成合成内容标识办法》同步施行，在监管维度建立审核、监测、追责的闭环机制，在技术维度规范标识体系并预留创新空间。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">中国社科院法学所研究员 支振锋：特别要提出的是这种生成式人工智能治理，我们一系列国家法律法规，还有规范性文件以及标准，一个很重要的治理目的是以规范促发展，通过治理和规范去净化互联网市场，净化网络信息内容、生态空间，从而促进人工智能特别是生成式人工智能，它的规范运用和健康良性的发展，实际上还是通过规范去促进发展的。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">另外，据专家介绍，高质量数据对生成式人工智能产业发展至关重要，基于错误或不完整的数据作出的决策可能偏离正确方向，不符合法规要求的数据可能带来法律风险。如果未经标识的内容进入公开数据领域，被大模型用于预训练会出现数据污染等情况，最终会导致技术可信度的大大降低，而“标识办法”的出台，正是从源头破解这一困局的关键举措。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">张吉豫：对于人工智能企业去选择相应的训练数据能够起到很好的指引，特别是所有的这样的生成内容都需要进行隐式的标识，这些标识是很便捷的，可以被机器检测的，就可以很便捷的去筛选哪些可能的内容是机器产生的，也便于他们去进行训练的选择。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">有利于公平竞争 减少侵权诈骗等问题\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">据专家介绍，在《人工智能生成合成内容标识办法》以及强制性的国家标准同步施行的情况下，标识的存在也会对大规模搬运人工智能直接生成内容的行为进行拦截，让内容平台有了更明确的合规要求和管理程序，有利于降低总体的审核运营成本。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">据法学专家介绍，《人工智能生成合成内容标识办法》的施行有利于公平竞争和有效管理。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">首先，强制“标识”有利于保护传统内容产业，能在一定程度上降低AI技术带来的冲击；\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">另外，通过标识可以追溯AI生成内容的来源，减少因AI生成内容引发的侵权、诈骗等问题，维护个人隐私和财产安全。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">支振锋：这些生成人工智能服务的使用者，比如说自媒体账号、博主，他们要使用这些服务的话，他们自己也有合规的要求，特别是一些MCN机构，一些自媒体账号，它其实背后有公司在运营，他们都有这样的合规义务，都要按照这个办法，还有国家标准去做生成合成内容的标识。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">据法学专家介绍，“标识办法”与《互联网信息服务深度合成管理规定》《生成式人工智能服务管理暂行办法》等法规充分衔接。其中，服务提供者在履行算法备案、安全评估等手续时，相关材料可能涉及标识的技术细节，如水印算法、元数据格式、服务商编码等，这些信息能够帮助执法部门在发现违法内容时快速识别内容来源，追踪到服务提供者或者用户。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】央视新闻客户端 \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.hrloo.com\u002Fnews\u002F347748.html\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">https:\u002F\u002Fwww.hrloo.com\u002Fnews\u002F347748.html\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fb5a96d19a3914ce59d531a9f49cd0cc7\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fthumbs\u002Fb5a96d19a3914ce59d531a9f49cd0cc7\u002FAI领域.jpg",0,1,42,"2025-09-11 18:16",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A34aa1abd-dc0e-42fe-b414-5eb3f91dfca4%3A0.wav?Expires=1757590546&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=HQqFTTUOif5QnwDBLD8synt3zcs%3D",10143556,"34aa1abd-dc0e-42fe-b414-5eb3f91dfca4","2025-09-11 18:12","Starting from tomorrow, AI-generated content must \"reveal its identity\".","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">The \"Artificial Intelligence Generated and Synthesized Content Identification Measures\" issued by the Cyberspace Administration of China and four other departments will be officially implemented starting September 1st. All AI-generated text, images, videos, and other content must \"reveal their identity\". Regarding who should manage these contents during the dissemination process and how to manage them?\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">The \"Artificial Intelligence Generated and Synthesized Content Identification Measures\" clearly states that artificial intelligence generated and synthesized content refers to information such as text, images, audio, video, and virtual scenes generated and synthesized using artificial intelligence technology. The identification of AI-generated and synthesized content includes explicit identification and implicit identification.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">When platforms put content on their service providers, they need to conduct reviews and verify the generated and synthesized content identification. They should add risk warnings for unmarked or suspected generated content, thereby blocking the spread of false information at the dissemination end.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Zhang Jiyu, Executive Director of the Institute of Future Law at Renmin University: The \"identification measures\" include three levels of detection. If a platform detects an implicit tag in the metadata, it can identify this content as AI-generated. However, if it only uses some algorithms to speculate that the content is AI-generated, it should be identified as potentially AI-generated.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Regarding situations where some original content is mistakenly identified as AI-generated due to technical limitations, resulting in infringement of creators' rights, legal experts pointed out that current algorithms still have issues with data bias and insufficient scenario adaptability. In cases where the error rate cannot be zero, corresponding safeguard mechanisms should be established.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Zhang Jiyu: After platforms implement graded identification, they should also establish graded handling methods. For example, for content clearly identified as AI-generated through implicit identification, stricter measures should be taken. However, for many contents that should be marked as potentially AI-generated, more convenient user feedback channels should be provided, and it is not advisable to impose serious penalties on users.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Legal experts suggest that first, online service providers should disclose algorithm principles and error sample data to enhance transparency. Second, they should set up appeal channels allowing users to submit evidence for review, thus ensuring the balance between regulation and creative freedom.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">Make generated content visible and traceable, promoting development through regulation\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">According to reporters, as of now, over 490 large models have been filed with the Cyberspace Administration of China, and over 240 large models have been registered with provincial cyberspace administrations. The user base of generative AI products in China has reached 230 million people. What impact will the implementation of the \"Identification Measures\" have on the industry, individual users, and related technological developments?\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">According to reporters, as the first combination of \"measures + strong standards\" in China's AI field, the mandatory national standard \"Cybersecurity Technology Artificial Intelligence Generated and Synthesized Content Identification Method\" will come into effect on September 1st along with the \"Artificial Intelligence Generated and Synthesized Content Identification Measures\", establishing a closed-loop mechanism for review, monitoring, and accountability in the regulatory dimension, and standardizing the identification system while reserving space for innovation in the technical dimension.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Zhi Zhenfeng, researcher at the Institute of Law, Chinese Academy of Social Sciences: It is particularly worth noting that the governance of generative AI, our series of national laws, regulations, and standards, has an important governance purpose of promoting development through regulation. By governing and regulating, we can purify the internet market and online information content and ecological space, thus promoting the standardized and healthy development of AI, especially generative AI. In fact, it is still about promoting development through regulation.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">In addition, according to experts, high-quality data is crucial for the development of generative AI. Decisions made based on incorrect or incomplete data may deviate from the correct direction, and data that does not meet regulatory requirements may bring legal risks. If unmarked content enters the public data domain and is used by large models for pre-training, it may lead to data pollution, ultimately greatly reducing the credibility of the technology. The introduction of the \"Identification Measures\" is a key measure to solve this problem at its source.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Zhang Jiyu: For AI companies, choosing appropriate training data can provide good guidance. Especially all such generated content needs to be implicitly marked. These marks are very convenient and can be detected by machines, making it easy to screen which content might be machine-generated, and facilitating their selection for training.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">Benefits for fair competition, reducing issues like infringement and fraud\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">According to experts, when the \"Artificial Intelligence Generated and Synthesized Content Identification Measures\" and the mandatory national standard are implemented simultaneously, the existence of the identification will also intercept large-scale copying of directly generated AI content, giving content platforms more explicit compliance requirements and management procedures, which is beneficial for reducing overall review and operational costs.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">According to legal experts, the implementation of the \"Artificial Intelligence Generated and Synthesized Content Identification Measures\" is beneficial for fair competition and effective management.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Firstly, the mandatory \"identification\" helps protect traditional content industries and can reduce the impact of AI technology to some extent;\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Additionally, by identifying the source of AI-generated content, it can reduce issues such as infringement and fraud caused by AI-generated content, protecting personal privacy and property security.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">Zhi Zhenfeng: Users of these generative AI services, such as self-media accounts and bloggers, have compliance requirements when using these services. Especially for some MCN agencies and self-media accounts, which are actually operated by companies, they have such compliance obligations and must follow this measure and national standards to identify generated and synthesized content.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan class=\"ql-lineHeight-1-75\" style=\"font-size: 18px;\">According to legal experts, the \"Identification Measures\" are well connected with other regulations such as the \"Internet Information Services Deep Synthesis Management Regulations\" and the \"Interim Measures for the Management of Generative AI Services\". When service providers fulfill procedures such as algorithm filing and security assessment, relevant materials may involve technical details of the identification, such as watermarking algorithms, metadata formats, and service provider coding. This information can help law enforcement agencies quickly identify content sources when discovering illegal content and trace it back to service providers or users.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">[News Source] CCTV News Client \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.hrloo.com\u002Fnews\u002F347748.html\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">https:\u002F\u002Fwww.hrloo.com\u002Fnews\u002F347748.html\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is reposted by this site to provide readers with more information and news. The content does not constitute investment or consumption advice. If there are any doubts about the facts of the article, please verify with the relevant parties. The views expressed in the article are not the views of this site and are for reference only.）\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A4c8a7d5a-3e42-49b5-9283-bcff04a37e14%3A0.wav?Expires=1774838473&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=ZPE%2Bz3hUElU3iDDlIL6%2F3l8S62g%3D","4c8a7d5a-3e42-49b5-9283-bcff04a37e14",13655444]