每周数据法律资讯 Data Law Weekly（20260713—20260719）

1.国家网信办发布第十八批深度合成服务算法备案公告

7月17日，国家网信办发布第十八批境内深度合成服务算法备案信息，本批次共计702个算法通过备案。公告明确，依据《互联网信息服务深度合成管理规定》第十九条规定，具有舆论属性或者社会动员能力的深度合成服务提供者，应当按照《互联网信息服务算法推荐管理规定》履行备案和变更、注销备案手续。深度合成服务技术支持者应当参照履行备案和变更、注销备案手续。尚未履行备案手续的深度合成服务提供者和技术支持者应尽快申请备案。

【参见：https://www.cac.gov.cn/2026-07/17/c_1786032856662750.htm?sessionid=】
 

CAC Releases the 18th Batch of Deep Synthesis Service Algorithm Filing Information

On July 17, the CAC released the 18th batch of domestic deep synthesis service algorithm filing information, a total of 702 algorithms were approved in this batch. The announcement clarifies that, pursuant to Article 19 of the Provisions on “the Administration of Deep Synthesis of Internet Information Services”, deep synthesis service providers with public opinion attributes or social mobilization capabilities shall complete filing procedures and procedures for filing changes or cancellations in accordance with “the Provisions on the Administration of Algorithm Recommendation for Internet Information Services”. Technical supporters of deep synthesis services shall perform filing procedures and procedures for filing changes or cancellations by reference. Deep synthesis service providers and technical supporters that have not yet completed filing procedures shall submit filing applications as soon as possible.

[Reference:

https://www.cac.gov.cn/2026-07/17/c_1786032856662750.htm?sessionid=]

2.安徽省通信管理局对11款逾期未整改的侵害用户权益APP实施下架处置

7月17日，安徽省通信管理局发布通报，针对2026年第三批专项整治中限期整改的APP开展复核，11款应用逾期未完成对侵害用户权益问题项的整改（如违规收集个人信息、超范围收集个人信息、APP频繁自启动和关联启动、过度索取权限等），监管要求各应用商店在通报发出后立即对该批APP全部下架。

【参见：

https://mp.weixin.qq.com/s/SlxzVy_sO3QABOJplILnWA】

Anhui Communications Administration Removes 11 Apps for Failure to Complete Rectification

On July 17, the Anhui Communications Administration issued a notice regarding the review of apps subject to rectification requirements under the third special campaign for 2026. Eleven apps failed to complete rectification within the prescribed period for issues infringing upon user rights and interests, including illegal collection of personal information, excessive collection of personal information, frequent automatic startup and associated startup, and excessive requests for permissions. The regulator required all application stores to immediately remove the relevant apps upon issuance of the notice.

[Reference:

https://mp.weixin.qq.com/s/SlxzVy_sO3QABOJplILnWA]

3.全国网安标委就《数据提供、委托处理、共同处理实施指南》征求意见

7月15日，全国网安标委就国家标准《数据安全技术数据提供、委托处理、共同处理实施指南》征求意见稿公开征求意见。本标准提出了数据对外提供、委托处理、共同处理的判定规则和责任划分，给出了数据处理活动的通用指南和具体场景实施指南。

【参见：

https://www.tc260.org.cn/portal/suggestion-detail/fc2b2c81ab0a4b37b59635b8349b05e8】

National Technical Committee 260 on Cybersecurity of SAC Seeks Public Comments on “the Guidelines for Data Provision, Entrusted Processing and Joint Processing”

On July 15, the National Technical Committee 260 on Cybersecurity of SAC released the draft national standard “Data Security Technology-Implementation Guidelines for Data Provision, Entrusted Processing and Joint Processing” for public comments. This standard establishes criteria for determining data provision, outsourced processing, and joint processing, as well as the allocation of responsibilities, and provides general guidelines for data processing activities along with implementation guidelines for specific scenarios.

[Reference:

https://www.tc260.org.cn/portal/suggestion-detail/fc2b2c81ab0a4b37b59635b8349b05e8]

4.全国网安标委就《网络安全技术人工智能应用安全分类分级方法》征求意见

7月15日，全国网安标委就国家标准《网络安全技术人工智能应用安全分类分级方法》征求意见稿公开征求意见。征求意见稿提出建立人工智能应用安全分类分级方法，适用于人工智能应用开发、运营等全生命周期安全活动。在分类方面，从应用场景属性、应用任务属性、应用智能化能力三个维度分析安全风险与类型；在分级方面，从影响严重程度和发生概率两方面综合研判各类安全风险级别，并按照就高从严原则进行综合定级。同时，标准提出分类分级宜参考行业领域细化标准，并根据应用属性和技术发展动态评价风险，定期审核更新安全级别，为人工智能应用安全管理提供参考。

【参见：

https://www.tc260.org.cn/portal/suggestion-detail/c8801e1c45954e098355a334d703a003】

National Technical Committee 260 on Cybersecurity of SAC Seeks Public Comments on “the Cybersecurity Technology-Method for Security Classification and Grading of Artificial Intelligence Applications”

On July 15,the National Technical Committee 260 on Cybersecurity of SAC released the draft national standard “the Cybersecurity Technology-Method for Security Classification and Grading of Artificial Intelligence Applications” for public comments. The draft proposes a method for security classification and grading of artificial intelligence applications, applicable to security activities throughout the lifecycle of AI application development and operation. For classification, the draft analyzes security risks and categories from three dimensions: application scenario attributes, application task attributes, and AI capability level. For grading, it evaluates security risk levels based on both the severity of impact and probability of occurrence, and determines the overall security level according to the principle of applying the higher and stricter standard. The draft further proposes that classification and grading should refer to sector-specific standards, dynamically evaluate risks based on application characteristics and technological development, and regularly review and update security levels, providing guidance for AI application security management.

[Reference:

https://www.tc260.org.cn/portal/suggestion-detail/c8801e1c45954e098355a334d703a003]

5.国家网信办发布7款提供手机端侧生成式人工智能服务已备案信息的公告

7月15日，为促进生成式人工智能服务创新发展和规范应用，网信部门会同有关部门按照《生成式人工智能服务管理暂行办法》要求，有序开展生成式人工智能服务备案工作，并公告新增7款提供手机端侧生成式人工智能服务的备案信息，包括“Apple智能”以及“华为小艺”等AI大模型，备案信息涵盖属地、备案单位、备案编号、适用手机终端等内容。

【参见：

https://mp.weixin.qq.com/s/5MTWh4pWVAlL71RQbU-Udg】

CAC Releases Filing Information for Seven On-Device Generative Artificial Intelligence Services

On July 15, to promote the innovative development and regulated application of generative artificial intelligence services, cyberspace authorities, together with relevant departments, have carried out generative AI service filing activities in an orderly manner in accordance with “the Interim Measures for the Management of Generative Artificial Intelligence Services”. CAC announced the filing information of seven newly added services providing on-device generative artificial intelligence services for mobile phones, including “Apple Intelligence” and Huawei’s “Xiaoyi” AI large models. The filing information includes the jurisdiction, filing entity, filing number, and applicable mobile phone terminals.

[Reference:

https://mp.weixin.qq.com/s/5MTWh4pWVAlL71RQbU-Udg]

6.安徽省通信管理局通报2026年第四批侵害用户权益APP

7月14日，安徽省通信管理局通报2026年第四批侵害用户权益APP。经近期拨测检查，发现20款APP存在超范围收集个人信息、违规收集使用个人信息、频繁索取权限、频繁自启动和关联启动等问题，并已向相关企业下达责令改正通知书，要求限期整改。截至目前，仍有10款APP未完成整改，相关企业应于2026年7月17日前落实整改要求，逾期未整改的，监管将依法依规组织开展相关处置工作。

【参见：

https://mp.weixin.qq.com/s/DweafW_lUkjBlt4FMpuncA】

Anhui Communications Administration Reports the Fourth Batch of Apps Infringing User Rights and Interests in 2026

On July 14, the Anhui Communications Administration issued a notice on the fourth batch of apps found to infringe user rights and interests in 2026. Through recent inspection and testing, the authority identified 20 apps involving issues such as excessive collection of personal information, unlawful collection and use of personal information, frequent requests for permissions, and frequent automatic startup and associated startup. The authority has issued rectification notices to the relevant enterprises, requiring them to complete corrections within the prescribed period. As of now, 10 apps have not completed rectification. The relevant enterprises shall implement corrective measures by July 17, 2026; otherwise, regulatory authorities will take further actions in accordance with laws and regulations.

[Reference:

https://mp.weixin.qq.com/s/DweafW_lUkjBlt4FMpuncA]

1.法国CNIL发布员工活动监控合规指引

近日，法国国家信息与自由委员会（CNIL）发布员工活动监控指引。CNIL指出，雇主有权基于工作管理和安全需要监督员工，但监控措施不得过度，并应同时满足目的正当、必要且适度，依法咨询员工代表机构，以及事先充分告知员工等要求。持续监控或使用键盘记录软件通常被认为不符合比例原则。

【参见：

https://www.cnil.fr/fr/controle-de-lactivite-des-personnes-employees】

French CNIL Issues Guidance on Employee Monitoring

Recently, the French Data Protection Authority (CNIL) issued guidance on employee monitoring. Employers may monitor employees for legitimate management and security purposes, but monitoring must not be excessive. Measures must be justified and proportionate, subject to prior consultation with employee representatives where required, and disclosed to employees in advance. Continuous monitoring and keylogging are generally considered disproportionate.

[Reference:

https://www.cnil.fr/fr/controle-de-lactivite-des-personnes-employees]