每周数据法律资讯 Data Law Weekly（20260504—20260510）

一

国内资讯（Domestic news）

1.工信部启动人工智能科技伦理审查与服务先导计划

5月9日，工业和信息化部印发通知，正式启动人工智能科技伦理审查与服务先导计划，先导计划部署了四项重点任务：一是细化省级人工智能科技伦理审查制度规范，健全城市科技伦理协同治理工作机制。二是指导各类创新主体建设人工智能科技伦理委员会，探索建设人工智能科技伦理审查与服务中心。三是开展人工智能科技伦理审查实践，组织开展高风险人工智能科技活动科技伦理专家复核，推动人工智能伦理风险评估、委员会建设等相关标准研制与验证。四是构建部省市三级联动敏捷治理网络，健全审查情况通报机制，建立人工智能科技伦理风险信息报送机制，完善风险预警与推送机制。

【参见：

https://mp.weixin.qq.com/s/NugYgQnXMhaayR8DuA3eGw】

MIIT Launches Pilot Program for AI Technology Ethics Review and Services

On May 9, MIIT issued a notice formally launching the Pilot Program for AI Technology Ethics Review and Services, setting out four key tasks: Firstly, to refine provincial-level regulations on AI technology ethics review systems and improve the collaborative governance mechanism for technology ethics at the city level. Secondly, to guide various innovation entities in establishing AI technology ethics committees and explore the establishment of AI technology ethics review and service centres. Third, to conduct practical AI ethics reviews, organise expert reviews of high-risk AI activities, and promote the development and validation of relevant standards for AI ethics risk assessment and committee establishment. Fourth, to establish a three-tiered (Ministry, provincial and municipal) agile governance network, improve the mechanism for reporting review findings, establish a mechanism for reporting AI ethics risk information, and refine risk early warning and notification mechanisms.

[Reference: 

https://mp.weixin.qq.com/s/NugYgQnXMhaayR8DuA3eGw]

2.网信办印发《智能体规范应用与创新发展实施意见》

5月8日，国家网信办印发《智能体规范应用与创新发展实施意见》，明确安全可控、规范有序、创新驱动、应用牵引四项原则。意见指出根据应用场景和潜在影响，审慎稳妥开展智能体分级治理。对于敏感领域及重点行业，由网信部门联合行业主管部门确定开放场景，根据相关法律法规、监管要求和安全防护标准，实行备案、检测、问题产品召回等管理措施。

【参见：

https://mp.weixin.qq.com/s/_b4zG-3_NH1CV7WpyLefgw】

CAC Issues “Implementation Opinions on the Standardised Application and Innovative Development of Intelligent Agents”

On May 8, the Cyberspace Administration of China(CAC) issued the “Implementation Opinions on the Standardised Application and Innovative Development of Intelligent Agents”, setting out four guiding principles: security and controllability, standardisation and orderliness, innovation-driven development, and application-led progress. The Opinions state that the classification and governance of intelligent agents should be carried out prudently and steadily based on application scenarios and potential impacts. For sensitive sectors and key industries, cyberspace administration departments will collaborate with relevant industry authorities to identify permitted application scenarios. In accordance with relevant laws, regulations, regulatory requirements, and security standards, management measures such as filing, testing, and recall of problematic products will be implemented.

[Reference:

https://mp.weixin.qq.com/s/_b4zG-3_NH1CV7WpyLefgw]
 

3.《生活服务类平台算法负面清单（试行）》实施取得初步成效

5月8日，中央网信办通报《生活服务类平台算法负面清单（试行）》实施进展。各地督促外卖、网约车、货运、网购、在线旅游等重点平台对照清单开展自查自纠，已实施优化改进措施63项，承诺遵守算法要求139项，限期推进125项，重点优化订单分配、时间预估、安全保障、收入抽成、动态定价、申诉处理等算法机制，推动就近派单、防疲劳、降低抽成、规范流量分配和提升算法透明度，在提升公平性、合理性方面取得初步成效。

【参见：

https://mp.weixin.qq.com/s/Ealhry1rGLqKWTzArOk9HQ】

Initial Success Achieved in the Implementation of the “Negative List of Algorithms for Lifestyle Service Platforms (Trial)”

On May 8, the CAC reported on the progress of implementing the “Negative List of Algorithms for Lifestyle Service Platforms (Trial)”. Local authorities have urged key platforms—including food delivery, ride-hailing, freight, online shopping and online travel services—to conduct self-inspections and rectifications in accordance with the list. To date, 63 optimisation and improvement measures have been implemented, 139 commitments to comply with algorithmic requirements have been made, and 125 measures are being advanced within specified timeframes. The focus has been on optimising algorithmic mechanisms such as order allocation, time estimation, safety safeguards, commission rates, dynamic pricing and complaint handling. Efforts have been made to promote localised order dispatch, prevent driver fatigue, reduce commission rates, standardise traffic distribution and enhance algorithmic transparency, achieving initial results in improving fairness and rationality.

[Reference:

https://mp.weixin.qq.com/s/Ealhry1rGLqKWTzArOk9HQ]

4.网信办就《消费类网联摄像头网络安全标识实施规则》及相关标准公开征求意见

5月8日，网信办发布通知，就《消费类网联摄像头网络安全标识实施规则》及《网络安全标识消费类网联摄像头安全要求》公开征求意见。规则明确消费类网联摄像头的网络安全标识对应的网络安全能力由低到高依次为基础级、增强级、领先级三个级别，相应的标识等级分别用一星、二星、三星表示，应重点检测物理硬件、系统软件、网络通信、数据和个人信息保护等安全能力，并要求产品完成备案后方可使用网络安全标识。意见稿明确备案流程、检测要求、有效期及违规处理机制，要求产品生产者对备案材料真实性负责，对伪造检测结果、冒用标识、虚假宣传等行为依法依规处理，同时推动通过扫码查询检测报告、关键指标等信息，提升消费类网联摄像头网络安全透明度和用户识别能力。

【参见：

https://mp.weixin.qq.com/s/EJe8jkD_QdmiI3q-shrOSg】

CAC Seeks Public Comments on the “Implementation Rules for Cybersecurity Labels for Consumer-Grade Network-Connected Cameras” and Related Standards

On May 8, the CAC issued a notice soliciting public comments on the “Implementation Rules for Cybersecurity Labels for Consumer-Grade Network-Connected Cameras” and the “Cybersecurity Label Security Requirements for Consumer-Grade Network-Connected Cameras”. The rules specify that the cybersecurity capabilities corresponding to the cybersecurity label for consumer-grade networked cameras are categorised into three levels from lowest to highest: Basic, Enhanced and Leading. The corresponding label levels are represented by one, two and three stars respectively. The rules stipulate that security capabilities such as physical hardware, system software, network communications, and data and personal information protection should be prioritised for testing, and require that products must complete the filing process before the cybersecurity label may be used. The draft specifies the filing procedures, testing requirements, validity periods and mechanisms for handling violations. It requires product manufacturers to be responsible for the authenticity of filing materials, and stipulates that acts such as falsifying test results, misusing the label or making false claims will be dealt with in accordance with laws and regulations. At the same time, it promotes the use of QR codes to access test reports and key performance indicators, thereby enhancing the transparency of cybersecurity for consumer-grade network-connected cameras and improving users’ ability to identify compliant products.

[Reference: 

https://mp.weixin.qq.com/s/EJe8jkD_QdmiI3q-shrOSg]

5.最高法发布人民法院依法惩治侵犯公民个人信息犯罪及关联犯罪典型案例

5月8日，最高法发布人民法院依法惩治侵犯公民个人信息犯罪及关联犯罪典型案例。本次发布的典型案例，聚焦行业内鬼泄露、非法获取医疗、出行等信息、学籍学历信息倒卖、网络“开盒”网暴、利用个人信息实施电信诈骗等突出类型，通过从源头打击、加强政企数据安全、防范网络暴力和严惩关联犯罪，实施数罪并罚和全链治理，强化法律威慑，保障公民权益和网络空间安全。

【参见：

https://mp.weixin.qq.com/s/UBLZw5XDfO6xLK9V4oXGUw】

Supreme People’s Court Publishes Typical Cases of People’s Courts Lawfully Punishing Crimes Involving the Infringement of Citizens’ Personal Information and Related Offences

On May 8,the Supreme People’s Court released a collection of typical cases involving the legal prosecution of crimes involving the infringement of citizens’ personal information and related offenses. The typical cases released this time focus on prominent types of offences, such as leaks by industry insiders, the illegal acquisition of medical and travel information, the trafficking of student registration and academic qualification data, online ‘unboxing’ and cyberbullying, and the use of personal information to carry out telecommunications fraud. By tackling these issues at source, strengthening data security for government and enterprises, preventing cyberbullying, and severely punishing related crimes, the authorities aim to implement cumulative sentencing for multiple offences and full-chain governance, thereby reinforcing the deterrent effect of the law and safeguarding citizens’ rights and the security of cyberspace.

[Reference: 

https://mp.weixin.qq.com/s/UBLZw5XDfO6xLK9V4oXGUw]

6.网信办发布第十七批深度合成服务算法备案信息

5月6日，国家网信办根据《互联网信息服务深度合成管理规定》，公开发布第十七批境内深度合成服务算法备案信息，相关主体可通过算法备案系统查询，未履行备案义务的深度合成服务提供者和技术支持者须尽快完成备案。

【参见：

https://www.cac.gov.cn/2026-05/06/c_1779809434590762.htm】

CAC Publishes the 17th Batch of Filing Information for Deep Synthesis Service Algorithms

On May 6, the CAC in accordance with the “Regulations on the Management of Deep Synthesis in Internet Information Services”, publicly released the 17th batch of domestic deep synthesis service algorithm filing information. Relevant entities may query this information via the algorithm filing system; deep synthesis service providers and technical supporters who have not fulfilled their filing obligations must complete the filing as soon as possible.

[Reference: 

https://www.cac.gov.cn/2026-05/06/c_1779809434590762.htm]

二

国外资讯（Overseas news）

1.欧盟推迟AI Act部分条款生效时间

5月8日，欧盟就修订《人工智能法案（AI Act）》达成简化措施协议，核心内容包括推迟高风险人工智能系统部分合规义务生效时间，以降低企业合规压力，同时保留风险分级监管框架；明确禁止未经同意生成儿童性虐待内容及裸照生成类应用，并要求违规系统在2026年12月前完成整改；消除机械设备领域合规重复要求，收紧安全组件认定范围，明确在保障条件下可处理个人数据以修正算法偏见；扩大中小企业豁免范围，优化通用人工智能系统执法流程。

【参见：

https://www.europarl.europa.eu/news/en/press-room/20260427IPR42011/ai-act-deal-on-simplification-measures-ban-on-nudifier-apps】

EU Postpones Entry into Force of Certain Provisions of the AI Act

On May 8, the EU reached an agreement on simplification measures for the revised AI Act. Key provisions include postponing the entry into force of certain compliance obligations for high-risk AI systems to alleviate compliance pressures on businesses, whilst retaining the risk-based regulatory framework; explicitly prohibiting the generation of child sexual abuse material and applications for generating nude images without consent, and requiring non-compliant systems to rectify issues by December 2026; duplicate compliance requirements in the machinery sector have been eliminated, the scope of safety component certification has been tightened, and it has been clarified that personal data may be processed to correct algorithmic bias under certain safeguards; the scope of exemptions for small and medium-sized enterprises has been expanded, and enforcement procedures for general artificial intelligence systems have been optimised.

[Reference: 

https://www.europarl.europa.eu/news/en/press-room/20260427IPR42011/ai-act-deal-on-simplification-measures-ban-on-nudifier-apps]

2.美国考虑对AI模型实施发布前审查

5月5日，据美国官员及知情人士透露，特朗普政府正考虑从此前主张减少干预、推动AI自由发展的立场转向加强监管，拟通过行政命令设立人工智能工作组，研究对新型AI模型的政府审查机制，并借鉴英国等国家的安全评估模式。白宫已与Anthropic、谷歌、OpenAI等企业沟通相关设想，可能在模型发布前引入安全与风险评估流程，但不会完全限制发布。

【参见：

https://www.reuters.com/world/white-house-considers-vetting-ai-models-before-they-are-released-nyt-reports-2026-05-04/?utm】

US Considers Pre-Release Review of AI Models

On May 5, according to US officials and informed sources, the Trump administration is considering a shift from its previous stance of advocating reduced intervention and promoting the free development of AI towards strengthened regulation. It plans to establish an artificial intelligence task force via executive order to study government review mechanisms for new AI models, drawing on the security assessment models of countries such as the UK. The White House has already discussed these proposals with companies such as Anthropic, Google and OpenAI. It may introduce safety and risk assessment processes prior to the release of models, but will not impose a complete ban on their release.

[Reference: 

https://www.reuters.com/world/white-house-considers-vetting-ai-models-before-they-are-released-nyt-reports-2026-05-04/?utm]

3.爱尔兰数据保护委员会对SHEIN启动合规调查

5月5日，爱尔兰数据保护委员会（DPC）宣布对SHEIN正式立案调查，以查明SHEIN爱尔兰公司是否违反欧盟通用数据保护条例，将欧盟/欧洲经济区用户的个人数据传输到中国，重点审查GDPR第5条、第13条及第5章规定的相关义务。

【参见：

https://www.dataprotection.ie/en/news-media/dpc-opens-inquiry-infinite-styles-services-co-ltd-shein-ireland】

Irish Data Protection Commission Launches Compliance Investigation into SHEIN

On May 5,the Irish Data Protection Commission (DPC) announced the formal opening of an investigation into SHEIN to determine whether SHEIN Ireland had breached the EU General Data Protection Regulation (GDPR) by transferring the personal data of EU/EEA users to China. The investigation will focus on the relevant obligations set out in Articles 5, 13 and Chapter 5 of the GDPR.

[Reference: 

https://www.dataprotection.ie/en/news-media/dpc-opens-inquiry-infinite-styles-services-co-ltd-shein-ireland]