Global AI Newsletter·Issue 25

On May 11, the General Office of the State Council issued theState Council 2026 Annual Legislative Work Plan (hereinafter the "Plan"). The Plan specifies that the 2026 legislative work should focus on building a more comprehensive socialist legal system with Chinese characteristics, strengthening legislation in key, emerging, and foreign-related areas. In particular, it emphasizes accelerating comprehensive legislation for the healthy development of AI, systematically improving legal frameworks governing common AI elements including data, computing power, algorithms, property rights, cybersecurity, and supply chain security, as well as standardizing key application scenarios. Concurrently, the Plan deploys the formulation of the Cybersecurity Graded Protection Regulation and the Online Trading Platform Supervision and Administration Regulation, the revision of the Internet Information Services Management Measures, and the advancement of preparatory legislative projects such as the Cybercrime Prevention and Control Law, comprehensively enhancing the rule of law in cyberspace.

Link: https://www.gov.cn/zhengce/content/202605/content_7068345.htm?sessionid=

2. CAC Reporting Center Releases Three Guides on Online Reporting, Rumor Debunking, and Infringement Reporting

On May 14, the Cyberspace Administration of China Reporting Center compiled and released three guides - the Online Reporting Guide, Online Rumor Debunking Guide, and Online Infringement Reporting Guide - aimed at systematically promoting knowledge of online reporting and rumor debunking basics, reporting channels, and specific procedures，, facilitating public participation in internet ecosystem governance. The three guides provide detailed instructions on reporting illegal and harmful information, identifying and reporting online rumors, and reporting procedures for online infringement involving personal privacy and reputation, offering pathway guidance for the public to protect their legitimate rights and interests.

Link: https://www.cac.gov.cn/2026-05/14/c_1780326306805270.htm?sessionid=

3. CAC Comprehensively Deploys and Advances Standardized Short Video Content Labeling

On May 12, the CAC summarized content labeling trial experiences from 12 platforms including Baidu, Pinduoduo, and Weibo, and guided platforms nationwide to comprehensively implement three tasks: first,standardizing short video content labels, requiring platforms to provide users with 6 categories of“"mandatory labels"”, with optional additional“"optional labels"”based on their circumstances. The“"mandatory labels"”include“"Contains fictitious or dramatized content,"”“"Contains AI-generated content,"”“"Contains marketing information,"”“"Content is reposted,"”“"Content represents personal opinion,"”and“"No label needed."”。Second,making content labeling a mandatory step in short video publishing, requiring creators to select one“"mandatory label"”before publishing.Third,strengthening review of newly added short video labels, conducting batch retrospective reviews of existing videos, supplementing or correcting improperly labeled content, and educating and warning relevant creators to ensure comprehensive labeling of all short video content.

Link:https://www.cac.gov.cn/2026-05/12/c_1780328273038196.htm

4. National Data Standards Committee Issues Plans for 15 Recommended National Standards/Normative TechnicalDocumentsin the Data Domain

On May 14, the National Data Standards Committee issued plans for 15 recommended national standards/normative technical documents in the data domain, including General Principles for Public Data Resource Development and Utilization, Technical Requirements for Data Interoperability Interface Management, High-Quality Dataset Construction Guide, High-Quality Dataset Data Synthesis Technical Requirements, High-Quality Dataset Construction and Operation Capability Maturity Model, Data Infrastructure Access Management, Data Infrastructure Regional and Industry Functional Node Technical Requirements, Trusted Data Space Data Value Assessment Framework, City-Wide Digital Transformation Urban Road Traffic Sensing Data General Requirements, High-Quality Dataset Quality Evaluation System Technical Requirements, High-Quality Dataset Embodied Intelligence Data Sources and Components, High-Quality Dataset Embodied Intelligence Simulation Synthetic Data Generation and Processing Specifications, High-Quality Dataset Embodied Intelligence Data Collection and Model Training Specifications for Training Bases, High-Quality Dataset Data Annotation Practitioner Competency Requirements, and Trusted Data Space Capability Building Requirements.

Link: https://mp.weixin.qq.com/s/GYDq9C25g5CrJyc0rUUK7Q

5. Beijing Municipal Bureau of Economy and Information Technology Issues Implementation Plan for AI-Empowered Industrial Internet High-Quality Development (2026-2028)

On May 13, the Beijing Municipal Bureau of Economy and Information Technology issued the Implementation Plan for AI-Empowered Industrial Internet High-Quality Development (2026-2028). The plan aims to create 100 high-quality industrial datasets, cultivate 50 intelligent solution providers, launch 100 high-level industrial AI agents, and build a batch of multi-agent collaborative application scenarios by 2028. The plan deploys key tasks across five areas: first, strengthening core technology development for industrial AI agents, accelerating intelligent transformation of industrial internet platforms, and building safety protection barriers; second, improving the high-quality industrial data supply system, enhancing data governance capabilities, and building a three-tier industrial AI agent dataset system; third, deepening full-chain scenario opening in key industries and Beijing-Tianjin-Hebei collaborative applications, focusing on high-precision industries such as integrated circuits, intelligent equipment, automotive, biomedicine, and aerospace, promoting AI agent empowerment across R&D design, production manufacturing, and quality control, creating a new "chain-style" digital-intelligent transformation model for industrial chains; fourth, building an integrated and open development ecosystem, promoting integration of industry, education, research, and application, and improving standards and certification systems; fifth, strengthening overall coordination mechanisms, ensuring policy alignment and promotion.

Link:https://www.ncsti.gov.cn/zcfg/zcwj/202605/t20260513_246570.html

6. Beijing CAC and Two Other Departments Formally Issue Data Export Negative List and Management Measures for China (Beijing) Pilot Free Trade Zone and National Comprehensive Demonstration Zone

On May11, the Beijing Municipal Internet Information Office, Beijing Municipal Bureau of Commerce, and Beijing Municipal Government Services and Data Management Bureau jointly issued the Management Measures (Trial) and accompanying Negative List (2025 Version) for the China (Beijing) Pilot Free Trade Zone and National Comprehensive Demonstration Zone for Expanding Opening-Up in the Services Sector.。

This policy adopts a“1+9”"1+9" framework design, comprising 1 set of“"Dual Zone"”management measures and 9 sector-specific negative lists, constructing a data export negative list management system applicable to Beijing's entire jurisdiction in a single release.The applicable scope has been expanded from the pilot free trade zone to the National Comprehensive Demonstration Zone for Expanding Opening-Up in the Services Sector (i.e., the entire Beijing municipality).. The lists cover 9 industry sectors, 67 business scenarios, and 612 data fields: first, the initial pilot free trade zone lists are directly converted to“"Dual Zone"”lists, covering 5 sectors including automotive, pharmaceuticals, civil aviation, retail, and AI; second, targeting Beijing's key industrial development directions in biomedicine, high-level autonomous driving, dual-airport international logistics hub, and international financial services, 4 new sectors are added including medical devices, autonomous driving (intelligent connected vehicles), trade logistics, and banking.。

The supporting management measures are optimized in three areas: first, improving the city-district two-level management system, with municipal departments leading and 17 districts (including the Economic Development Area) organizing implementation; second, application materials follow the minimum necessity principle, retaining only“"one form, one letter"”(the Application Form and Enterprise Commitment Letter), with all other materials eliminated; third, security supervision shifts from“"universal inspection"”to“"credit + risk"”tiered and classified precision supervision, implementing lightweight supervision for highly-rated enterprises.。

Link: https://mp.weixin.qq.com/s/hsif6NovRtMF3OmjCdk67A

7. Guangzhou AI Industry Development Office Issues Guangzhou AI Industry 2026 Work Priorities

On May 11, the Guangzhou AI Industry Development Office issued the Guangzhou AI Industry 2026 Work Priorities (hereinafter the "Work Priorities"). The Work Priorities propose 32 specific measures across eight areas including forging industrial innovation engines, implementing industrial factor strengthening, implementing intelligent upgrades in key sectors, and optimizing the industrial ecosystem, aiming to accelerate Guangzhou's development into a vertical model capital and AI application demonstration highland. Key measures include: building 2 AI industry common technology service platforms, cultivating 40 competitive vertical large models and 450 AI application scenarios; completing the establishment of an AI industrial investment fund in the first half of 2026; strengthening the tiered computing power layout of "city data centers + park computing centers"; improving data circulation and trading mechanisms, deepening public data authorized operations, accelerating the construction of trusted data spaces for cities, industries, and enterprises; releasing no fewer than 50 application scenarios for "AI + Healthcare"; and accelerating the issuance of the Guangzhou AI Development Promotion Regulation and its implementation rules.

Link:https://finance.sina.com.cn/tech/roll/2026-05-11/doc-inhxppih2221319.shtml

8. Heilongjiang Provincial Communications Administration Solicits Public Comments on the Heilongjiang Province“"15th Five-Year Plan"”Information and Communications Industry Development Plan (Draft for Public Comment)

On May 11, the Heilongjiang Provincial Communications Administration issued a notice soliciting public comments on the Heilongjiang Province "15th Five-Year Plan" Information and Communications Industry Development Plan (Draft for Public Comment) (hereinafter the "Draft"). The Draft is divided into sections on development foundations and situation analysis, overall requirements, main tasks, and safeguard measures. Development targets specify that by 2027, the layout of new information infrastructure will be basically complete, with 5G coverage in cities, counties, towns, and scenic areas, and gigabit fiber broadband achieving comprehensive urban-rural coverage with "10-gigabit backbone" coverage in key industrial parks; by 2030, an intelligent and comprehensive information infrastructure system that is high-speed, ubiquitous, space-ground integrated, cloud-network converged, intelligently agile, green and low-carbon, and secure and controllable will be basically established. Key tasks include enhancing gigabit optical network and 5G deep coverage, advancing computing infrastructure layout, deepening industrial internet applications, and strengthening network and data security capabilities.

Link:https://hljca.miit.gov.cn/bsfw/gsgg/art/2026/art_795bca686be54c7aa36e5632785fd6c6.html

9. Hunan Province High-Quality Dataset Supply-Demand Matchmaking Conference Held in Changsha

On May 13, the Hunan Province High-Quality Dataset Supply-Demand Matchmaking Conference, themed "Model-Data Resonance, Intelligent Empowerment for Xiaoxiang," was held at the Hunan Big Data Exchange. Through three segments of on-site observation, supply-demand exchange, and thematic forums, the conference built a precision matchmaking platform for deep collaboration between high-quality datasets and AI, achieving intended cooperation of RMB 61.07 million on-site, and releasing nearly RMB 100 million in data annotation orders and 2,012 AI data talent positions. At the conference, the Hunan Provincial Department of Industry and Information Technology released the 2026 Hunan Province AI Application Demand and Supply Capability "Two Lists," listing 254 application demands and 518 supply capabilities; the Hunan Provincial Data Bureau simultaneously released the province's first batch of 25 high-quality dataset pioneer trial lists, covering tourism, transportation, healthcare, agriculture, industry, energy, and other sectors. The conference also announced the first batch of Changsha data annotation industry-education integration talent training bases.

Link:https://www.nda.gov.cn/sjj/swdt/dfdt/0514/20260514131647721507645_pc.html

10. Sichuan Conducts World Telecommunication Day Events, Releasing“"Digital Lifeline"”Core Capabilities and Innovation Institutions

On May 15, Sichuan Province, as part of its World Telecommunication Day events, released multiple core capabilities of China Telecom's“"Digital Lifeline,"”covering quantum security, low-altitude economy, green computing power, data elements, and network security; meanwhile, several innovation research institutions jointly established by China Telecom with the Sichuan Surveying, Mapping and Geoinformation Bureau, the Sichuan Earthquake Administration, and other entities - including the Emergency Housing and Construction Industry Research Institute Southwest Innovation Center and earthquake large model research - were officially inaugurated, accelerating the transition of digital resilience building from technological innovation to industrial implementation.

Link:https://gzw.sc.gov.cn/scsgzw/CU231003/2026/5/15/26c00db6f47f4421be06297e986458dd.shtml

1.2 Law Enforcement and Judicial Updates

1. First Nationwide AI-Written“"Product Recommendation Notes"”Case Decided

According to CCTV.com on May 12, the Hangzhou Intermediate People's Court of Zhejiang Province disclosed a typical case involving AI batch-generated product recommendation notes on a social platform, marking the first nationwide unfair competition case involving AI-written“"product recommendation notes."”Previously, the defendant operated an AI writing tool that could provide users with one-click generation services for product recommendation posts, travel guides, and other content in the style of the social platform, and induced users to publish the content on the platform. The platform operator argued that the defendant's actions disrupted the platform's authentic product recommendation content ecosystem and filed suit. The court held that providing targeted generative AI services for the social platform disrupted the plaintiff platform's content ecosystem centered on authentic experiences, constituting unfair competition, and ordered the defendant to compensate the plaintiff for economic losses and reasonable expenses totaling RMB 100,000.

Link:https://news.cctv.com/2026/05/12/ARTI8WYlR9mPd6KyPj3GhncE260512.shtml

2. “"First Nationwide AI Large Model Defamation Case"”Upheld on Final Appeal

According to China Jiangsu Net on May 12,“"the first nationwide AI large model defamation case"”was concluded on final appeal at the Nanjing Intermediate Court. Previously, Nanjing lawyer Li Xiaoliang searched his name and title on the Baidu mobile app and website and discovered that the“"AI Smart Answer"”displayed that he had“"been sentenced to three years imprisonment for the crime of arson,"”accompanied by a photo of him in a lawyer's robe. After multiple complaints through lawyer's letters, hotlines, and email channels went unresolved, Li filed suit against Baidu, demanding deletion of all false information, a top-of-page apology on Baidu's homepage for 8 months, compensation of RMB 150,000 for lost income, and damages for emotional distress. The Nanjing Jiangbei New District People's Court found Baidu liable for defamation at first instance and ordered a written apology. Baidu appealed. The Nanjing Intermediate People's Court recently dismissed the appeal and upheld the original judgment. After the final judgment took effect, Baidu failed to fulfill its apology obligation within the statutory period, and Li has submitted a compulsory execution application to the court.

Link:https://jsnews.jschina.com.cn/jsyw/202605/t20260512_s6a033847e4b04d95bbc5ad9a.shtml

3. Hangzhou Internet Court Releases 2025 Annual“"Judicial Index"”

According to China News Service on May 13, the Hangzhou Internet Court released its 2025 annual internet development“"Judicial Index."”。“"Judicial Index"”is based on analysis of judicial data from over 70,000 internet dispute cases accumulated in 2025 and historical judicial data, scoring 23.15% higher than 2018, the highest value in recent years. In the AI-related field, the“"Judicial Index"”shows that the number of related cases in 2025 increased by 32% compared to 2024, involving numerous industries and spanning both business operations and daily life. Hangzhou Internet Court Vice President Hua Lijun pointed out that generative AI has achieved leapfrog adoption at the technology, user, and industry levels, with its accompanying risks escalating from technical hazards to systemic social governance challenges. AI deepfakes have proliferated, with face-swapping and voice-changing technologies being used to create fake advertisements, commit telecom fraud, and forge public figure images, seriously undermining the social trust baseline.

Link:https://www.chinanews.com.cn/sh/2026/05-13/10620549.shtml

4. Shanghai Launches“"Qinglang - Rectifying AI Application Abuses"”Special Campaign

On May 15, Shanghai launched the "Qinglang - Rectifying AI Application Abuses" special campaign in two phases. The first phase is the "Qinglang - Typical Violations in AI Application Services" special governance action, focusing on rectifying issues such as failure to fulfill large model filing and registration obligations as required, insufficient security review capabilities, large model training corpus security, AI data poisoning, and inadequate implementation of generated/synthetic content labeling, strengthening source governance of AI technology. The second phase is the "Qinglang - Rectifying AI Information Content Abuses" special campaign, targeting the use of AI to generate "digital slop," create and publish false information, disseminate violent and vulgar content, impersonate others, infringe on minors' rights, and engage in online astroturfing, resolutely cleaning up illegal content and lawfully handling violating accounts, MCN organizations, and website platforms.

Link: https://mp.weixin.qq.com/s/O-7qJoUbrhsRtibfC71n9Q

5. Ministry of Public Security Detects 41 Mobile Apps Illegally Collecting and Using Personal Information

5May14, the Ministry of Public Security Computer Information System Security Product Quality Supervision and Inspection Center detected 41 mobile apps illegally collecting and using personal information, including 20 apps failing to disclose personal information collection and use rules; 23 apps collecting and using personal information without user consent; 1 app failing to fully and accurately inform users about personal information collection and use, or where the stated purpose, method, and scope of collection differed from actual practice; 1 app failing to list the types, purposes, methods, and recipient names and contact information for personal information provided to third parties; 6 apps failing to simultaneously inform users of the purpose when requesting permission to collect personal information or requesting sensitive personal information; 1 app forcing users to consent to collection of non-essential personal information; 7 apps collecting personal information beyond necessary scope, collecting location, contacts, SMS, and other information in unrelated scenarios; 2 apps failing to provide effective account cancellation functions; 3 apps failing to establish and publish effective personal information security complaint and reporting channels; 4 apps failing to provide users with convenient channels to exercise rights to correct, delete, or refuse processing of personal information; and 1 app with frequent or unexpected pop-up advertisements affecting normal use.

Link: https://mp.weixin.qq.com/s/OTdm7faodcYhGkAJpRHQQQ

1.3 International CooperationUpdates

1.MIIT and APEC Member Economies Hold 72nd APEC Telecommunications Working Group Meeting in Shanghai

On May 11, the 72nd APEC Telecommunications Working Group meeting was held in Shanghai, with MIIT Vice Minister attending the opening ceremony and delegations from APEC member economies' digital and ICT authorities participating. The meeting, centered on the 2026 APEC“"Open, Innovation, Cooperation"”three priority areas, conducted policy exchanges and pragmatic consultations on advancing Asia-Pacific digital infrastructure connectivity, digital economy cooperation, and AI governance, with China committing to deepen digital economy collaboration with all parties and elevate Asia-Pacific digital cooperation to new heights.

Link:https://www.miit.gov.cn/xwfb/bldhd/art/2026/art_22463e9dad584eff979f9ecac9da1a60.html

1.4 Research Updates

1.UIBE National Institute for Open Economy Releases Blue Book on Opening Up: Beijing Opening-Up Development Report (2025)

On May 13, the University of International Business and Economics National Institute for Open Economy released the Blue Book on Opening Up: Beijing Opening-Up Development Report (2025). The blue book dedicates a chapter to analyzing Beijing's digital economy internationalization, noting that Beijing has positioned itself on new digital opening tracks, leading nationally in AI large model filings (132, accounting for nearly 35% of the national total) and computing infrastructure; in cross-border data flows, it has introduced negative lists and“"green channel"”measures to facilitate digital trade; it recommends further advancing core digital infrastructure including 5G and big data centers, deepening smart city digital management, and continuing to attract international digital economy enterprises and talent.

Link:https://finance.sina.com.cn/jjxw/2026-05-13/doc-inhxtvfk9280546.shtml

2. CAICT Releases Southeast Asia Regional Computing Center Service Provider Analysis Report (2026)

On May 13, CAICT released the Southeast Asia Regional Computing Center Service Provider Analysis Report (2026), analyzing the 2025 regional computing market size, regional layout, and competitive landscape by country, assessing industrial development opportunities and core challenges in resources, policies, and talent, evaluating the difficulty of Southeast Asian computing overseas expansion, building a systematic evaluation of regional service providers, projecting 2030 industry trends, and proposing targeted development recommendations at policy and industry levels.

Link: https://www.caict.ac.cn/kxyj/qwfb/ztbg/202605/t20260513_722939.htm

3. “"Data Factor ×"”City Tour First Stop Event Officially Launched in Nanjing

On May 13, the "Data Factor ×" City Tour first stop event, co-hosted by the National Data Development Research Institute and the Jiangsu Provincial Data Bureau, was launched in Nanjing, Jiangsu. The event officially established the Data Factor Marketization and Value Realization Collaboration Network, and released multiple important results including the Jiangsu Data Factor Quality Scenario List and the Data Scenario Demand List.

Link: https://mp.weixin.qq.com/s/ZbhIrzXCL5c2r6Lu-RjUEQ

4. Data Circulation Service Institution Mutual Recognition Initiative Advancement Group Holds Semi-Annual Meeting inJiangsu

On May 13, the Data Circulation Service Institution Mutual Recognition Initiative Advancement Group held its semi-annual meeting at the Jiangsu International Data Port Conference Center. Over 100 representatives from major national data exchanges (centers), leading industry data circulation service platform enterprises, and data providers attended. Jiang Jiang, Vice President of the National Data Development Research Institute, attended and delivered remarks. The meeting established five specialized working groups - general standards, data product mutual recognition, high-quality datasets, data assetization, and self-regulatory standards - focusing on key areas including standardization, product mutual recognition, assetization pathways, and industry self-regulation, with clear priority tasks defined. Elections for the advancement group leader and specialized working group leaders were also conducted during the meeting.

Link:https://mp.weixin.qq.com/s/eUuomlelnJ1g71sn5SmB3Q

1.5 Industry Updates

1. ZTE Appears at Latin America M360 Summit, Sharing AI and Network Convergence Solutions

On May 15, ZTE shared its insights on the bilateral empowerment value between AI and telecommunications networks, AI-Native network innovation practices, and full-scenario intelligent solutions at the 2026 Latin America M360 Conference hosted by GSMA. The conference focused on commercial monetization, network resilience, and inclusive growth, exploring core topics including AI-empowered networks, next-generation connectivity strategies, cybersecurity, data infrastructure, and digital skills,promotingtangible economic and social valuecreation,helping Latin American operators achieve strategic upgrades.

Link:https://gu.qq.com/resources/shy/news/detail-v2/index.html?t=1#/index?_tentrees_trans=0&id=SN202605151752209740f5bf

2. 7th Shanghai Information Consumption Festival Officially Launches, Focusing on Smart Computing Cloud and Metaverse

On May 17, the 7th Shanghai Information Consumption Festival officially launched, running through June 17, themed“"Shanghai Smart New Ecosystem - Digital Sharing New Consumption."”This festival focuses on three core sectors: smart computing cloud, metaverse, and digital content, with 36 themed activities planned. The smart computing cloud sector features 10 key activities with over 30 participating companies, distributing over RMB 3 million in benefits around cloud computing, AI, and AI agents. The metaverse sector presents an LBE large-space touring exhibition showcasing multiple local benchmark content offerings.

Link:https://news.qq.com/rain/a/20260516A03NVZ00?adChannelId=sh

3. Intelligent IP WAN Application Advancement Action and China Tour Officially Launched

On May 16, the 2026 World Telecommunication and Information Society Day Conference was held in Wuhan. At the conference, the China AI Industry Development Alliance (AIIA) Information and Communications AI Innovation Advancement Committee, CAICT, together with the China Communications Standards Association and the three major telecom operators and other industry partners, jointly launched the "Intelligent IP WAN (AI WAN) Application Advancement" Action and China Tour. Based on existing innovative practices of basic telecommunications enterprises, the action focuses on four directions: reducing industry computing costs through elastic and efficient enterprise computing access networks, promoting smart home computing service innovation through flexible and fast edge computing access networks, supporting distributed AI training and inference through high-speed lossless new metropolitan area networks, and facilitating efficient data element circulation and management through secure and reliable high-speed data networks.

Link:https://www.stcn.com/article/detail/3912371.html

II. International Governance Updates

2.1 Policy and Legislative Updates

1. G7 Jointly Releases Guide on Minimum Elements for AI Software Bill of Materials

On May 12, the U.S. Cybersecurity and Infrastructure Security Agency (CISA), Germany's Federal Office for Information Security (BSI), Italy's National Cybersecurity Agency (ACN), France's National Agency for the Security of Information Systems (ANSSI), Canada's Communications Security Establishment (CSE), the UK National Cyber Security Centre (NCSC), and Japan's National Center of Incident Readiness and Strategy for Cybersecurity (NCO) jointly released the Guide on Minimum Elements for AI Software Bill of Materials (Software Bill of Materials for AI – Minimum Elements), aiming to define the minimum elements of AI software bills of materials to enhance system transparency and cybersecurity.

The guide divides AI software bills of materials into seven key data clusters:（1）Metadata. Metadata records basic information about the software bill of materials itself, such as author, version, data format, and signing status.（2）System-level attributes. System-level attributes primarily record global information about the AI system, including system models, component inventories, data flows, and data use planning.（3）Model cluster. The model cluster focuses on the identity structure and training status of AI algorithm models, enabling lifecycle protection and traceability, including model name, description, and hash values.（4）Dataset cluster.The dataset cluster records key information about data used throughout the model lifecycle,including dataset name, content, statistical characteristics, and sensitivity.（5）Infrastructure cluster. The infrastructure cluster encompasses the physical and virtual environments required for system operation, including firmware, frameworks, and runtime environments.（6）Security attributes cluster.The security attributes clusterfocuses on cybersecurity management and control measures for AI systems, including security controls, vulnerabilities, and access controls.（7）Performance cluster. The performance cluster quantitatively evaluates the security level and operational performance of AI systems and models to support security monitoring, primarily including system uptime, time-to-resolution, and latency metrics.

Link:https://www.bsi.bund.de/SharedDocs/Downloads/EN/BSI/KI/SBOM-for-AI_minimum-elements.pdf?__blob=publicationFile&v=4

2. U.S. Senate Banking Committee Approves Revised Digital Asset Market Clarity Act

On May 12, the U.S.Senate Banking Committee released the full revised text of the Digital Asset Market Clarity Act (CLARITY Act), which was approved by the committee on May 14 with a vote of 15-9. The nine-chapter bill's core content includes: delineating SEC and CFTC jurisdiction over digital assets; establishing decentralization test standards (with five indicators - open source code, permissionless neutrality, distributed network meaning no single coordinating entity holds more than 49%, autonomous distributed ledger, and economic independence), with tokens from networks passing the test classified as commodities under CFTC regulation, while those failing are deemed“"subsidiary assets"”subject to disclosure and registration obligations; standardizing digital asset exchange registration, DeFi protocol compliance requirements, custodian qualifications, and a four-year transitional certification window for initial offerings; and retaining compromise provisions on stablecoin yield. This bill represents the first systematic digital asset market structure legislation in the United States, and if ultimately passed, would provide a legal framework for U.S. non-stablecoin digital asset regulation.

Link:https://www.techflowpost.com/en-US/article/31567

3. UK Government Issues Policy Guidance on AI, Open Source Code, and Vulnerability Risks in the Public Sector

On May 14, the UK Government Digital Service (GDS) and the Department for Science, Innovation and Technology (DSIT) jointly released the guide on AI, Open Source Code, and Vulnerability Risks in the Public Sector（AI, open code and vulnerability risk in the public sector）. The guide notes that the primary risk of systems being exploited stems from their own security flaws and insufficient remediation capabilities, rather than source code publicity; while AI can accelerate attacks, the additional risk from code visibility is relatively limited. The guide proposes four core recommendations: first, ensuring public systems meet minimum operational standards including clear ownership, security by design, automated dependency management, and trusted remediation capabilities; second, adhering to the“"open by default"”principle, avoiding a shift to“"private by default"”due to AI threats; third, code that must be closed should be accompanied by clear threat model documentation, with exceptions having time limits and periodic review; fourth, strengthening remediation capabilities by setting patch timelines and automating vulnerability management. The guide also warns that making code private should not substitute for substantive investment in security capabilities, as privatization may create a false sense of security and undermine cross-departmental collaboration and external scrutiny.

Link:https://www.gov.uk/guidance/ai-open-code-and-vulnerability-risk-in-the-public-sector

4. Illinois Advances AI Safety Measures Act

On May 11,the U.S. state ofIllinois is advancing the Artificial Intelligence Safety Measures Act. The bill primarily targets“"frontier models,"”“"frontier developers,"”and“"large frontier developers,"”establishing safety governance rules. Under the bill,“"frontier models"”are defined as foundation models whose training process uses computing power exceeding 10^26 integer or floating-point operations, including the original training process and subsequent fine-tuning, reinforcement learning, or other substantive modifications.

Key provisions include: first, requiring large frontier developers to develop, implement, and publicly disclose a frontier AI framework explaining how they handle catastrophic risk management, transparency, and cybersecurity; second, requiring developers to publish transparency reports when deploying new frontier models or making significant modifications to existing ones, disclosing model release dates, supported languages, output modalities, intended uses, and other information; third, requiring large frontier developers to report critical safety incidents to the Illinois Emergency Management and Homeland Security Office, and prohibiting materially false or misleading statements regarding catastrophic risk management of frontier models.

Link:https://legiscan.com/IL/research/SB3312/2025

5. Kazakhstan Signs Decree on Measures to Introduce AI into Secondary Education System

According to Kazinform on May 12, Kazakhstan's PresidentKassym-Jomart Tokayevsigned the decreeOn Measures to Introduce Artificial Intelligence into the Secondary Education System of the Republic of Kazakhstan.Under the decree,the government must broadly introduceartificial intelligenceinto the secondary education system by July 1, 2026, submit specific recommendations for pilot projects to narrow the education quality gap between urban and rural schools by June 1, ensure schools are equipped with stable high-speed internet by August 1, and develop application standards and teacher continuing professional development plans by September 1. Standards will cover academic integrity norms, norms for the use ofartificial intelligencein educational content generation and learning outcome assessment, while also strengthening student personal data protection.

Link:https://cn.inform.kz/news/lejiangrengongnengzhongxuedefa-a14c14db

6. Japan Digital Agency Solicits Public Comments on Two Draft Orders Related to Individual Number System

On May 11, Japan's Digital Agency, pursuant to the Act on the Use of Numbers to Identify Specific Individuals in Administrative Procedures (Individual Number Act), launched public comment periods for two draft supporting orders aimed at clarifying the scope of individual number usage and rules for providing specific personal information, and improving data provision standards for government information network systems. The draft orders include:first,，a draft order defining the affairs determined by competent ministerial orders in the Act's appendices;second,，a draft order, based on the revision of Article 19, Item 8 of the Act, on the provision of specific personal information. The public comment period runs from May 11toJune 9, 2026.

Link:https://www.digital.go.jp/news/42bba26a-8167-49f3-9e0f-12fcc619cba7

2.2 Law Enforcement and Judicial Updates

1. Florida Mass Shooting Victims' Families Sue OpenAI

According to Reuters on May 11, families of victims of a mass shooting in Florida have filed a lawsuit against OpenAI. The shooter, Phoenix Ikner, killed two people at a school in Tallahassee, Florida. The court alleges the shooter received assistance from ChatGPT in planning the attack. The victims' families claim Ikner used information provided by ChatGPT in conversations over the preceding months to plan and execute the attack, with conversations discussing mass shootings, the lethality of Ikner's weapon, and the busiest times at the FSU student union. The chatbot did not flag or escalate these conversations, and thefamiliesconsider ChatGPT a co-conspirator in the shooting.OOpenAI maintains that ChatGPT provided answers to fact-based questions, that the information is widely available through public internet channels, that it did not encourage or facilitate illegal or harmful activities, and that it should not bear liability in this case.

Link:https://www.reuters.com/legal/government/family-florida-mass-shooting-victim-sues-openai-us-court-2026-05-11/

2. Parents of Man Who Died from Drug Overdose Sue OpenAI

According to Reuters on May 12, the parents of a man who died last year from an accidental drug overdose have filed a lawsuit in California court against OpenAI and its founder and CEO, alleging that the man was guided by ChatGPT to take a dangerous combination of multiple substances. The parents claim that ChatGPT told Nelson how to obtain illegal substances, suggested which drug he should take next, and made recommendations based on Nelson's reported related experiences, and that the combination of these drugs and alcohol caused the man's death.

Link:https://www.reuters.com/legal/litigation/openai-faces-lawsuit-california-court-claiming-chatbot-gave-advice-that-led-2026-05-12

3. Supreme Court of Victoria, Australia Issues Court User Practice Guide and Judicial Officers AI Usage Guidelines

On May 14, the Supreme Court of Victoria, Australia issued a Court User Practice Guide and Judicial Officers AI Usage Guidelines. The Court User Practice Guide states that court users are responsible for the content of documents submitted to the court regardless of whether AI was used, and that submitting court documents containing inaccurate information may result in cost orders. Lawyers relying on unverified AI outputs in court documents may also be referred to the Victorian Legal Services Board and Commissioner and may face personal cost limitations. The Judicial Officers AI Usage Guidelines state that generative AI must not be used for judicial decision-making, but court-approved AI tools may be used to assist judges and court staff in organizing and searching case materials, generating summaries and timelines from case materials, assisting legal research, and proofreading - these uses cannot substitute for reading and hearing evidence and submissions, or fact-finding required in judicial decision-making.

Link:https://www.supremecourt.vic.gov.au/news/new-practice-note-and-judicial-guidelines-on-the-use-of-artificial-intelligence

4. UK Information Commissioner's Office Reminds Organizations to Address AI-Driven Cyber Threats

On May 14, the UK Information Commissioner's Office published a blog reminding organizations to take measures against AI-driven cyber threats. The blog proposes five practical steps: first, understanding the main AI-driven risks, including AI-enhanced phishing and deepfake social engineering; second, maintaining basic、multi-layered defense; third, limiting access points and applying the principle of least privilege; fourth, enhancing detection, monitoring, and incident impact capabilities through regular vulnerability scanning and penetration testing to identify weaknesses, maintaining and regularly testing incident response plans; fifth, protecting personal data through data minimization, storage limitations, and data audits.

Link:https://ico.org.uk/about-the-ico/media-centre/news-and-blogs/2026/05/five-steps-to-protect-your-organisation-from-ai-powered-cyber-threats

5. TikTok Challenges EU "Gatekeeper" Designation at European Court of Justice

According to Reuters on May 12, TikTok has filed an appeal with the Court of Justice of the European Union (CJEU) seeking to overturn its designation as a "gatekeeper" under the Digital Markets Act (DMA). Under the DMA,“"gatekeepers" are large online platforms with over 45 million monthly active users and entrenched market positions, required to comply with strict obligations on interoperability, data sharing, and non-self-preferencing, with violations punishable by fines of up to 10% of global annual revenue. TikTok was designated as a "gatekeeper" by the European Commission in September 2023, alongside Google, Meta, Apple, and other tech giants. A lower court had previously dismissed TikTok's challenge.

TikTok's lawyers argued at the hearing that 70-80% of the company's users simultaneously use multiple social platforms, and that its valuation primarily derives from Asian operations with weak ties to the European market, failing to meet DMA requirements for "entrenched market position." EU Commission lawyers countered that even with multi-homing, certain user groups may still be locked into the TikTok ecosystem. The court will issue its ruling in the coming months. Meta has also filed similar challenges regarding its Messenger and Marketplace "gatekeeper" designations.

Link:https://www.reuters.com/sustainability/boards-policy-regulation/tiktok-challenges-eu-gatekeeper-status-europes-top-court-2026-05-12/

6. NetflixAccused of Deceptive Data Practices and Children's Privacy Misleading

On May 11,theTexas Attorney General announced a lawsuit against Netflix, alleging deceptive practices in user data collection, advertising operations, and children's privacy protection, in alleged violation of the Texas Deceptive Trade Practices Act (DTPA).

The lawsuit claims Netflix has long promoted itself as an ad-free, privacy-focused streaming service, but has in fact established an advertising-supported subscription model and uses user viewing behavior, device information, and interaction records to support ad targeting and commercial analytics. The suit also specifically addresses children's data protection, arguing that Netflix has not adequately protected minors' rights regarding children's profiles, autoplay features, and personalized recommendations. The Texas Attorney General requests the court to enjoin the relevant data collection and disclosure practices, require Netflix to disable autoplay by default on children's accounts, and seeks injunctive relief and civil penalties.

Link: https://www.texasattorneygeneral.gov/news/releases/attorney-general-ken-paxton-sues-netflix-spying-texas-kids-and-consumers-illegally-collecting-users

7. U.S. Texas Restricts LG from Collecting"Viewing Data"

through Automatic Content Recognition TechnologyOn May 11,the

Texas Attorney General reached a settlement with LG Electronics U.S.A., Inc. (LG) requiring it to adjust smart TV viewing data collection practices. The case primarily involves Automatic Content Recognition (ACR) technology, which identifies content on TV screens or audio/video feeds to analyze what programs, apps, or external device content users are watching, using the data for advertising and content analytics.

Under the settlement, LG must provide consumers with clear, prominent explanations before using viewing data, and state at the top of relevant agreements that if users consent, LG will use ACR technology to identify and analyze content on devices. LG must also update viewing information agreements on smart TVs and websites, and must not continue collecting viewing data without clear disclosure and valid consent.

8. U.S. FTC Reminds Major Technology Platforms to Comply with TAKE IT DOWN Act

On May 11, the U.S. Federal Trade Commission (FTC) sent letters to Amazon, Alphabet, Apple, Meta, Microsoft, Reddit, TikTok, X, and over a dozen other technology companies, reminding them to comply with the Take It Down Act (TIDA) by May 19.

The law primarily targets the non-consensual distribution of intimate images, videos, and related deepfake content, requiring covered platforms to establish victim removal mechanisms. The FTC's concurrent compliance guidance states that platforms should delete relevant content and known identical copies within 48 hours of receiving a valid request; complaint mechanisms should be clear, easily accessible, and allow victims or their authorized representatives to submit requests.

Link: https://www.ftc.gov/news-events/news/press-releases/2026/05/ftc-chairman-ferguson-advises-companies-comply-take-it-down-act?utm_source=.com

9. U.S. FTC and ShutterstockReachSettlement over Alleged Illegal Subscription and Cancellation Practices

On May 13, the U.S. Federal Trade Commission（FTC）announced that Shutterstock has agreed to pay $35 million to resolve alleged illegal subscription and cancellation practices.

The FTC alleged that Shutterstock, in its digital image, video, and music stock subscription services, failed to adequately disclose auto-renewal terms, cancellation fees, and other material charges, did not obtain explicit informed consent from consumers, and complicated cancellation processes. Under the proposed order, in addition to the settlement payment, Shutterstock must cease misrepresenting subscription terms, clearly disclose material information before charging, obtain explicit consumer consent, and provide simple cancellation methods.

Link:https://www.ftc.gov/news-events/news/press-releases/2026/05/shutterstock-pay-35-million-settle-ftc-allegations-over-illegal-subscription-cancellation-practices?utm_source.com

10. Belgian Data Protection Authority Issues Three Fines Totaling Over 370,000 Euros

On May 16, the Litigation Chamber of the Belgian Data Protection Authority (APD) published three penalty decisions totaling approximately 376,000 euros before the departure of its president Hielke Hijmans.

First, Walloon water utility company SWDE was fined 86,000 euros for GDPR violations in its call center recordings, including: insufficient transparency to callers (requiring internet access for complete information), failure to sign subcontracting agreements for nearly 5 years, partially illegal test recordings, and failure to ensure recordings were deleted within one month (a prerequisite for the call center exception).

Second, fintech company Isabel SA was fined 120,000 euros. The company operates the identity verification service TruliUs but incorrectly classified itself as a“"data processor"”rather than a“"data controller,"”resulting in failure to respond to user access requests and excessive collection of personal information (including identity card photos), violating the accountability principle.

Third, a major technology company was fined approximately 176,000 euros for failing to delete a former employee's work email six months after departure. The company had no legitimate interest in retaining the mailbox long-term, and failed to inform the data subject and correspondents about data processing or respect their access and deletion rights.

The APD simultaneously issued remediation orders. All parties may appeal within 30 days.

Link: https://www.autoriteprotectiondonnees.be/citoyen/la-chambre-contentieuse-inflige-3-amendes

2.3 International CooperationUpdates

1. Vietnamese and Austrian Experts Hold Talks on AI and Cybersecurity Cooperation

According to Voice of Vietnam on May 12,a delegation from the Vietnam National Cybersecurity Association (NCA,，NCA) and representatives of Vietnam's Ministry of Public Security held a working meeting in Vienna with the Austrian Institute of Technology (AIT,）AIT), engaging in in-depth discussions on strengthening cooperation in cybersecurity, data, and AI. During the meeting, the NCA proposed establishing a strategic partnership with AIT, with key cooperation areas including protecting critical digital infrastructure, building secure data and trustworthy digital ecosystems, applying AI and data analytics to cybersecurity, high-quality personnel training, expert exchanges, and jointly hosting annual international forums and workshops. The Vietnamese side also proposed expanding connections between technology enterprises of both countries, promoting technology transfer and joint projects in cybersecurity, data, and AI. AIT experts presented advanced technology applications including large-scale data analytics, AI-driven communication technologies, intelligent border monitoring, post-quantum security, and digital identity management.

Link:https://vovworld.vn/news/vietnamese-austrian-experts-discuss-ai-cybersecurity-cooperation-2434275.vov5#

2. Third UN Virtual Worlds Day Held in Geneva

On May 11-12, the Third UN Virtual Worlds Day was held at the International Telecommunication Union (ITU) headquarters in Geneva, Switzerland, themed "AI, Spatial Intelligence, and the Citiverse: Shaping Trusted Digital Futures." This event was jointly organized by the ITU in collaboration with nearly 20 UN agencies and international organizations including UNDP, UN-Habitat, UNICEF, the World Bank, and WIPO, in cooperation with global city centers.

The event brought together ambassadors, ministers, city leaders, UN agency representatives, and industry pioneers for high-level dialogue on the future of AI-powered cities and communities, including an ambassadors' roundtable. Topics focused on the integration of frontier technologies such as AI, immersive virtual environments, and spatial intelligence in urban governance and public services, exploring emerging concepts such as the AI-empowered "Citiverse," and examining how to translate the Global Digital Compact into trustworthy, inclusive, and people-centered outcomes for cities and communities.

Link:https://www.itu.int/un-virtual-worlds-day/2026/

3. 2026 Vietnam Automation Technology Expo and China-Vietnam AI Industry Cooperation Forum Held in Hanoi

According to ASEAN Social,on May 14-16, the 2026 Vietnam Automation Technology Expo (Automation Technology Expo 2026) was held in Hanoi, Vietnam.。The concurrent Second China-ASEAN (Vietnam) AI Industry Development Summit Forum attracted over 400 guests from multiple countries, including leaders from Vietnam's Ministry of Science and Technology, National Science Association, and the Ministry of Science and Technology's Productivity Research Institute. During the forum, the Vietnam Automation Association, China-ASEAN Community of Shared Future Promotion Association, Global Unicorn 500 Engineering, Gree Electric Vietnam, and other organizations jointly released the China--Vietnam AI and Automation Cooperation Initiative（(hereinafter the "Hanoi Consensus").）The Hanoi Consensus, building on the Joint Statement issued following the meeting of the two countries' leaders in April 2026, aims to promote AI industrial ecosystem construction, jointly build digital economy cooperation zones, and facilitate intelligent manufacturing standards mutual recognition,，specifying that China and Vietnam will deepen pragmatic cooperation in areas including joint large model R&D, computing resource coordination, and industry standard mutual recognition.。The conference also hosted sessions focusing on green technology, AI-empowered low-carbon environmental protection, digital economy thematic discussions, and entrepreneur exchanges, showcasing achievements of AI in facilitating traditional industry transformation and upgrading.

Link:https://mp.weixin.qq.com/s/sEW1OroKBVW_tR85g2b2Mw

4. UNCTAD Convenes 9th Session of Intergovernmental Group of Experts on E-Commerce and Digital Economy

On May 11-13, the United Nations Conference on Trade and Development (UNCTAD) convened the 9th Session of the Intergovernmental Group of Experts on E-Commerce and the Digital Economy in Geneva.

This session focused on the theme of enhancing fiscal revenue capacity of developing countries in the context of e-commerce and digital trade. Discussions covered digital economy tax governance, the role of e-commerce platforms in tax administration, the support of digital solutions for tax authority capacity building, and developing country participation in digital trade rule-making. Related discussions involved taxation, digital economy measurement, e-commerce policy formulation, and platform economy governance capacity building, providing policy references for developing countries to improve digital economy regulation and data statistics systems.

Link: https://unctad.org/meeting/intergovernmental-group-experts-e-commerce-and-digital-economy-9th-session?utm_source.com

ꄴPrevious： null

ꄲNext： null