曼昆周报 | Ai×Web3 每周法律简报 (2026-06-10)

⚡ Week in Review: June 10–17, 2026

This edition covers critical AI and Web3 legal, regulatory, and enforcement developments across the United States, China, Hong Kong, and the EU during the week of June 10–17, 2026, with compliance actions for AI and Web3 companies. A comparatively quiet week — much of the cycle's momentum ran in late May and early June — so several items below remain in a procedural or implementation phase. Prepared for cross-border practice (New York & PRC).

I & Web3 Law — Key Developments

U.S. LEGISLATION — FEDERAL

June 10–17 — CLARITY Act awaits Senate floor time (Calendar No. 423).

The Digital Asset Market Clarity Act remained on the Senate Legislative Calendar this week after clearing the Senate Banking Committee 15–9 (May 14) and its June 1 placement. The bill would grant the CFTC exclusive jurisdiction over digital-commodity spot markets while preserving SEC jurisdiction over investment-contract assets, and adds a DeFi-protocol framework, stablecoin-yield limits, and a digital-commodity insolvency safe harbor. Obstacles persist: the 60-vote floor threshold, an unresolved ethics provision on government officials' crypto holdings, and roughly eight weeks of calendar before the summer recess — making the White House's aspirational July 4 signing increasingly tight. No floor vote had occurred as of this week.

Ongoing — GENIUS Act stablecoin rulemaking advances at the OCC.

Implementation of the GENIUS Act (enacted July 18, 2025) continues via the OCC's March 2, 2026 proposed rule governing issuance of payment stablecoins by entities under OCC jurisdiction. The framework requires 100% reserve backing with qualifying liquid assets and limits issuance to permitted payment stablecoin issuers. Issuers should confirm reserve auditing and public attestation readiness.

Ongoing — AI chip export enforcement exposure persists.

Following Commerce's June 1 re-clarification that the advanced-AI-chip ban reaches China-headquartered firms operating outside China, suppliers and distributors routing hardware through third-country intermediaries (Malaysia, Singapore, UAE) face presumptively non-compliant arrangements. Taiwan's separately reported (June 9–10) consideration of criminal penalties for unauthorised Blackwell-class exports to all of China would, if enacted, be the most expansive extension of the chip-control regime outside the U.S.

U.S. LEGISLATION — STATE

Pending — New York's three-bill AI package awaits Governor Hochul's signature.

The FAIR News Act (AI-content disclosure), the AI Training Data Transparency Act (A 6578, requiring a published dataset summary for any model offered in New York), and the Kids AI Chatbot Safety Bill (S 9051, prohibiting companion chatbots for minors) passed June 4–5 before sine die; Governor Hochul is expected to sign all three. Separately, New York's RAISE Act (frontier-model safety and 72-hour incident reporting) was signed December 19, 2025 and takes effect January 1, 2027.

Ongoing — California's ~30 AI bills advance in second-chamber committees.

California's remaining AI bills — covering AI liability, biometric data, automated decision systems, and AI in healthcare — are moving through second-chamber committees ahead of the July 2 recess, with sine die on August 31. The national chatbot wave continues, with roughly 78 active chatbot bills across 27 states.

Ongoing — Colorado's replacement ADS framework.

SB 26-189 (signed May 14) repealed the original, EU-modelled Colorado AI Act and replaced it with a narrower disclosure-and-rights framework for automated decision-making, effective January 1, 2027. Confirm clients are not still building to the repealed high-risk standard.

CHINA

This week — CAC opens a dedicated AI-misconduct reporting channel under Qinglang 2026.

The CAC Reporting Center launched a specialised hotline and online portal for AI-application misconduct as part of the "Qinglang Rectifying AI Application Chaos" special campaign, coordinated with the Ministry of Public Security. The channel accepts complaints across 14 violation categories, including failure to label AI-generated/synthetic content, AI-generated false or misleading information, and harmful generative content. This is the most concrete new enforcement step of the week: it gives the 2026 campaign a standing intake mechanism, and unregistered generative-AI services remain a primary target. The 2025 campaign removed 3,500+ AI products and 960,000+ pieces of illegal content.

July 15 deadline (~28 days) — Anthropomorphic AI Interactive Services Interim Measures.

The CAC's Interim Measures for the Management of Anthropomorphic AI Interactive Services (issued April 10, 2026) take effect July 15, 2026, governing AI companions, chatbots, and virtual assistants that simulate human personality. Key obligations: privacy compliance, safety testing, mandatory AI labelling, and a ban on companion features for minors. There is no grace period.

Ongoing — CAC/NDRC/MIIT AI Agents framework in implementation.

The joint Implementation Opinions on the Standardised Application and Innovative Development of Intelligent Agents (issued May 8, 2026) remain in active implementation — the first Chinese framework treating AI agents as a distinct category from generative AI/LLMs. The risk-tiered structure subjects healthcare, finance, transportation, judicial services, and public security to mandatory filing and joint-regulator recall mechanisms, while lower-risk consumer/productivity applications face lighter self-assessment.

EU

Ongoing — GPAI Code of Practice finalised; enforcement infrastructure operational ahead of Aug 2.

Building on the June 1 appointment of the 60-member Scientific Panel and the June 3 publication of the transparency Code of Practice on Marking and Labelling of AI-Generated Content, the General-Purpose AI Code of Practice has been finalised ahead of the August 2, 2026 commencement of the Commission's GPAI supervision and enforcement powers (documentation requests, evaluations, compliance measures, and fines). The Commission also advanced a tech-sovereignty package (June 3) to strengthen Europe's digital autonomy.

Confirmed effective December 2, 2026 — new prohibited practices.

The EU confirmed that AI systems generating non-consensual intimate imagery or CSAM will be prohibited from December 2, 2026, with fines up to €35 million or 7% of global turnover. A standardised visual "AI" label is required on AI-generated text, images, audio, and video, with a safe harbour for Code signatories.

HONG KONG (WEB3)

Ongoing — First stablecoin issuer licences move toward launch.

Under the Stablecoins Ordinance (effective August 2025), the HKMA's first batch of issuer licences — granted April 10 to HSBC and Anchorpoint Financial — are progressing to launch, with HSBC planning an HKD-denominated stablecoin in H2 2026. The HKMA received 36 formal applications and continues to advance despite Beijing's maintained mainland crypto prohibition, widening the two-track divergence. Mainland-nexus clients should maintain clear operational and legal firewalls between Hong Kong structures and the mainland.

CALL TO COMPLIANCE ACTIONS

What AI Companies Should Do Now

U.S. — Federal

Frontier-model developers — Trump AI EO: Assess exposure to the voluntary pre-release access framework and document which models may qualify as "covered frontier models" once the classified benchmark publishes. Audit AI tools that automate external-system access for inadvertent CFAA exposure, given the EO's DOJ enforcement priority on AI-assisted intrusion.
Training-data provenance: Maintain a documented due-diligence programme; identify and quarantine any data from piracy sources, unlicensed scraping of paywalled content, or content with AI-use restrictions. The ~$1.5B Bartz v. Anthropic settlement (~$3,100/work) is the current damages benchmark.

U.S. — New York

Training Data Transparency Act (A 6578): If you offer a generative-AI model to New York users, build a public-facing dataset-summary disclosure now, before Governor Hochul's signature makes the obligation imminent.
Kids AI Chatbot Safety Bill (S 9051): If you operate any companion/social/interactive chatbot accessible to under-18s in New York, plan immediate service restrictions or age verification — passage was unanimous.
FAIR News Act: Implement an AI-generated-content disclosure/labelling system for news distributed in New York.

China

Anthropomorphic AI — July 15 deadline (~28 days): Complete a gap analysis against the CAC Interim Measures now — privacy compliance, safety testing, mandatory AI labelling, and a ban on companion features for minors. No grace period.
Qinglang 2026 — registration audit: Verify all generative-AI services deployed in China hold current CAC registration; unregistered services are the primary enforcement target and a public reporting channel is now live.
AI Agents framework — high-risk filing: If deploying AI agents in healthcare, finance, transportation, judicial services, or public security, begin mandatory filing with the relevant sectoral authority in coordination with the CAC.

EU

GPAI compliance — enforcement live Aug 2: Ensure GPAI model documentation, capability evaluations, and systemic-risk assessments are complete; consider signing the Code of Practice as the fastest path to Article 50 transparency compliance.

What Web3 Companies Should Do Now

U.S.

CLARITY Act — token-portfolio mapping (urgent): Map your token portfolio against the Act's CFTC/SEC jurisdictional divide; the decentralisation threshold is the critical variable. Engage digital-assets counsel ahead of any floor vote.
GENIUS Act — stablecoin reserve compliance: If you issue or plan to issue a U.S. stablecoin, confirm 100% reserve backing with qualifying liquid assets per the OCC proposal, and implement reserve auditing and public attestation.
AI-chip / infrastructure supply chain: If hardware is procured through third-country intermediaries, review sourcing chains immediately; Commerce's re-clarification makes such arrangements presumptively non-compliant.

Hong Kong

Stablecoin licensing: With HSBC and Anchorpoint licensed and launches approaching, ensure reserve documentation, governance, and AML/KYC programmes are current; assess eligibility now if not yet applied..

This briefing is prepared for informational purposes only and does not constitute legal advice.If you are seeking AI legal governance, or Web3 & digital assets compliance, feel free to reach out to our lawyers. Many thanks!

Mankun Law PLLC

⚡ 本周回顾：2026年6月10日至17日

本期涵盖2026年6月10日至17日当周，美国、中国、香港及欧盟在AI与Web3()领域的关键法律、监管及执法动态，并附AI与Web3企业的合规行动建议。本周相对平静——本轮立法与司法的主要动能集中在5月下旬至6月初——因此下文多项事项仍处于程序推进或落地实施阶段。本周报面向跨境实务（美国纽约州与中国）。

本期周刊梳理过去一周AI与Web3领域最值得关注的监管、立法及司法动态，并结合最新趋势提出企业合规建议，供读者参考。

AI 与 Web3 法律——重要动态

美国法院

6月10日–17日（持续进行）——OpenAI 多地区合并诉讼（MDL()）：证据开示扩大，累计AI知识产权索赔突破500亿美元。合并审理的 In re OpenAI Copyright Litigation 案（纽约南区联邦法院），由 Sidney Stein 法官主审、整合16起诉讼，本周仍处于积极的证据开示阶段。继1月5日命令强制提供2,000万条匿名化 ChatGPT 对话日志、3月9日进一步追加7,800万条及1,000万条日志后，原告律师正在审查这些数据，以寻找复制及“记忆”受版权保护文本的证据。此次数据提供规模在AI知识产权诉讼中前所未有；截至6月1日的法院文件，所有在审AI版权及相关知识产权案件的累计主张赔偿金额已超过500亿美元。
6月11日——Bartz 诉 Anthropic() 案：和解分配金额完成计算；最终批准仍待裁定。本周迎来6月11日这一节点，和解管理人据此计算15亿美元和解金的分配方案，款项预计于6月或之后开始发放（视是否提起上诉而定）。本周是6月11日的里程碑，和解管理人将计算15亿美元和解协议下的分配，预计将在6月或更晚开始支付（取决于任何上诉）。法官Araceli Martínez-Olguín（案件重新分配）在听证会上没有签署最终命令。
持续进行——CNN 诉 Perplexity 案：首例广播新闻类AI版权诉讼推进。CNN 于5月29日提起诉讼，指控 Perplexity 系统性抓取并复制其受版权保护的新闻内容，案件已进入证据开示阶段。CNN 选择诉讼而非授权，标志着传统媒体的立场趋于强硬。

美国立法——联邦

6月10日–17日——CLARITY 法案等待参议院全院表决时间（议事日程第423号）。《数字资产市场清晰法案》（Digital Asset Market Clarity Act）本周仍列于参议院立法议事日程，此前已于5月14日以15比9获参议院银行委员会通过，并于6月1日列入日程。该法案将赋予商品期货交易委员会（CFTC()）对数字商品现货市场的专属管辖权，同时保留证券交易委员会（SEC）对投资合同类资产的管辖权，并新增DeFi()协议框架、稳定币收益限制以及数字商品破产安全港条款。障碍依然存在：全院60票通过门槛、关于政府官员持有加密资产的未决道德条款，以及夏季休会前仅剩约八周的议程时间——这使得白宫期望中的7月4日签署日期愈发紧张。截至本周，全院尚未进行表决。
持续进行——GENIUS 法案稳定币规则制定在 OCC 层面推进。《GENIUS 法案》（2025年7月18日颁布）的实施通过货币监理署（OCC）2026年3月2日发布的拟议规则继续推进，该规则规范受 OCC 管辖主体的支付型稳定币发行。框架要求以合格流动资产提供100%储备支持，并将发行限于获许可的支付型稳定币发行人。发行人应确认储备审计与公开鉴证的准备情况。
持续进行——AI芯片出口管制风险持续。继商务部6月1日重新澄清先进AI芯片禁令适用于在中国境外运营的中国总部企业后，通过第三国中介（马来西亚、新加坡、阿联酋）输送硬件的供应商和分销商面临被推定为不合规的安排。台湾另据报道（6月9日–10日）正在考虑对未经授权向中国全境出口 Blackwell 级芯片处以刑事处罚，若获通过，将成为美国以外对芯片管制制度最大范围的延伸。

美国立法——州层面

待签署——纽约州三项AI法案组合等待州长 Hochul 签署。《FAIR新闻法》（要求披露AI生成内容）、《AI训练数据透明法》（A 6578号，要求就在纽约州提供的任何模型公布数据集摘要）以及《儿童AI聊天机器人安全法》（S 9051号，禁止面向未成年人的陪伴型聊天机器人）已于6月4日–5日休会前通过；州长 Hochul 预计将签署上述三项法案。另外，纽约州《RAISE法案》（前沿模型安全及72小时事件报告）已于2025年12月19日签署，将于2027年1月1日生效。
持续进行——加州约30项AI法案在第二院委员会推进。加州其余AI法案——涵盖AI责任、生物识别数据、自动化决策系统及医疗AI——正在第二院委员会审议，须赶在7月2日休会前推进，本届会期于8月31日结束。全国性的聊天机器人立法浪潮持续，27个州约有78项在审聊天机器人法案。
持续进行——科罗拉多州替代性自动化决策框架。SB 26-189号法案（5月14日签署）废止了原本仿照欧盟模式的《科罗拉多AI法》，代之以更为狭窄的、针对自动化决策的披露与权利框架，于2027年1月1日生效。请确认客户不再按照已废止的高风险标准进行合规建设。

中国

本周——国家网信办在“清朗2026”专项行动下开通AI乱象专门举报渠道。中央网信办举报中心开通了针对AI应用违法违规行为的专门热线和在线平台，将“清朗·整治AI应用乱象”专项行动落到实处，该行动由公安部协同开展。举报渠道受理涵盖14类违规行为的公众举报，包括未对AI生成或合成内容进行标识、利用AI生成虚假或误导性信息，以及生成有害内容。这是本周最为具体的新执法举措：它为2026年专项行动提供了常态化受理机制，未经登记备案的生成式AI服务仍是重点打击对象。2025年专项行动下架AI产品逾3,500款，清理违法内容逾96万条。
7月15日生效（约28天）——拟人化AI交互服务管理暂行办法。网信办《人工智能拟人化交互服务管理暂行办法》（2026年4月10日发布）将于2026年7月15日生效，规范模拟人类人格的AI陪伴、聊天机器人及虚拟助手。主要义务包括：隐私合规、安全测试、强制性AI标识，以及禁止面向未成年人的陪伴功能。不设过渡宽限期。
持续进行——网信办/发改委/工信部AI智能体框架进入实施阶段。三部门联合发布的《关于智能体规范应用和创新发展的实施意见》（2026年5月8日发布）仍在积极实施——这是中国首个将AI智能体作为区别于生成式AI/大语言模型的独立类别加以规范的框架。该分级框架要求医疗、金融、交通、司法服务及公共安全领域履行强制备案并建立多部门联合召回机制，低风险的消费/生产力类应用则适用较为宽松的自评估。
欧盟持续进行——通用人工智能（GPAI）行为准则定稿；执法基础设施在8月2日前已就绪。在6月1日任命60名成员的科学小组、6月3日发布《AI生成内容标记与标识行为准则》的基础上，通用人工智能（GPAI）行为准则已定稿，为欧盟委员会自2026年8月2日起启动的GPAI监督与执法权力（要求提供文件、开展评估、要求合规措施及处以罚款）做好准备。委员会另于6月3日推出科技主权一揽子方案，以增强欧洲的数字自主能力。确认于2026年12月2日生效——新增禁止性行为。欧盟确认，自2026年12月2日起，将禁止生成非自愿私密影像（NCII）或儿童性虐待材料（CSAM）的AI系统，违者最高处以3,500万欧元或全球营业额7%的罚款。AI生成的文本、图像、音频和视频须加注标准化的“AI”可视标识，行为准则签署方可享有安全港。 香港（Web3）
持续进行——首批稳定币发行人牌照迈向落地运营。在《稳定币条例》（2025年8月生效）下，香港金管局首批发行人牌照——于4月10日授予汇丰银行（HSBC）和 Anchorpoint Financial——正推进至业务落地，汇丰计划在2026年下半年推出港元（HKD）计价稳定币。金管局共收到36份正式申请，并在北京维持内地加密货币禁令的背景下继续推进，使“两条轨道”的差异进一步扩大。涉及内地业务的客户应在香港主体与内地之间保持清晰的运营与法律隔离。

合规行动建议

AI企业当前应采取的行动

美国——联邦层面

前沿模型开发者——特朗普AI行政令： 评估在自愿性“发布前访问”框架下的风险敞口，并在机密基准发布后记录哪些模型可能符合“受涵盖前沿模型”的标准。鉴于该行政令将“AI协助的入侵”列为司法部执法重点，应审查会自动访问外部系统的AI工具，以排查潜在的《计算机欺诈与滥用法》（CFAA）风险。
训练数据来源审查：维护书面的尽职调查程序；识别并隔离来自盗版来源、未经授权抓取付费内容或带有AI使用限制的数据。约15亿美元的 Bartz 诉 Anthropic 和解（每部作品约3,100美元）是当前的赔偿基准。

美国——纽约州

训练数据透明法（A 6578）： 若您向纽约州用户提供生成式AI模型，应立即建立面向公众的数据集摘要披露，赶在州长 Hochul 签署使义务即将生效之前。
儿童AI聊天机器人安全法（S 9051）： 若您运营任何未成年人（18岁以下）在纽约州可访问的陪伴/社交/交互型聊天机器人，应规划立即的服务限制或年龄验证——该法案系全票通过。
FAIR新闻法： 为在纽约州分发的新闻建立AI生成内容的披露/标识系统。

中国

拟人化AI——7月15日截止（约28天）： 立即对照网信办暂行办法进行差距分析——隐私合规、安全测试、强制AI标识，以及禁止面向未成年人的陪伴功能。无宽限期。清朗2026——备案审查： 核实在中国部署的所有生成式AI服务均持有有效的网信办备案；未备案服务是重点打击对象，且公众举报渠道现已上线。AI智能体框架——高风险备案： 若在医疗、金融、交通、司法服务或公共安全领域部署AI智能体，应会同网信办向相关行业主管部门启动强制备案。

欧盟

GPAI合规——8月2日执法生效： 确保GPAI模型文件、能力评估及系统性风险评估完备；可考虑签署行为准则，作为满足第50条透明度合规的最快途径。

Web3企业当前应采取的行动

美国

CLARITY法案——代币组合梳理（紧急）： 对照该法案的CFTC/SEC管辖划分梳理您的代币组合；“去中心化门槛”是关键变量。在任何全院表决前应聘请数字资产律师。GENIUS法案——稳定币储备合规： 若您发行或计划发行美国稳定币，应按OCC拟议规则确认以合格流动资产提供100%储备支持，并落实储备审计与公开鉴证。芯片/基础设施供应链： 若通过第三国中介采购硬件，应立即审查采购链；商务部的重新澄清使此类安排被推定为不合规。

香港

稳定币牌照：随着汇丰和 Anchorpoint 获牌、业务即将落地，应确保储备文件、治理结构及反洗钱/了解你的客户（AML/KYC）程序均为最新；尚未申请的应立即评估资格。

中国

严格规避——内地业务： 任何与中国内地存在关联的产品须完全规避加密货币交易、稳定币发行、真实世界资产（RWA）通证化及Web3服务提供；香港主体须保持清晰的运营与法律隔离。

本简报仅供参考，不构成法律意见。如需AI法律治理或Web3与数字资产合规服务，欢迎联系我们的律师。感谢您的阅读！

曼昆周报 | Ai×Web3 每周法律简报 (2026-06-10)