US, China and EU take different paths on AI regulation — Eter News

Enriched Analysis 深度解析

US, China and EU take different paths on AI regulation美国、中国与欧盟采取不同的人工智能监管路径

The United States, China and the European Union are pursuing distinct approaches to governing artificial intelligence. Washington emphasizes voluntary industry commitments and executive action, China combines expanding rules with continued development, while the EU relies on its AI Act and digital-services oversight, balancing safety concerns against fears of limiting innovation.美国、中国和欧盟正采取不同的人工智能治理方式。美国强调企业自愿承诺与行政行动，中国在扩大监管的同时继续推动发展，欧盟则依靠《人工智能法案》和数字服务监管，在安全担忧与限制创新的风险之间寻求平衡。

AFP 7 Oct 2026 2026年10月7日

9090

Measures in the AI Action Plan《人工智能行动计划》中的措施数量

01The Story新闻事实

The United States, China and the European Union are pursuing distinct approaches to governing artificial intelligence. Washington emphasizes voluntary industry commitments and executive action, China combines expanding rules with continued development, while the EU relies on its AI Act and digital-services oversight, balancing safety concerns against fears of limiting innovation.美国、中国和欧盟正采取不同的人工智能治理方式。美国强调企业自愿承诺与行政行动，中国在扩大监管的同时继续推动发展，欧盟则依靠《人工智能法案》和数字服务监管，在安全担忧与限制创新的风险之间寻求平衡。

  • The United States has not enacted a federal AI law, while state-level AI laws are taking effect nationwide.美国尚未通过联邦人工智能法律，但各州人工智能法律正在全国范围内生效。
  • The White House’s July 2025 AI Action Plan listed more than 90 measures to accelerate development.白宫2025年7月公布的《人工智能行动计划》列出了90多项加速发展的措施。
  • China has rules covering AI use by children and AI companion bots, including a ban on creating digital clones of real people without consent.中国已制定涵盖儿童使用人工智能和人工智能陪伴机器人的规定，包括禁止未经同意创建真实人物的数字克隆体。
  • The EU adopted its AI Act in 2024 and later delayed rules covering high-risk systems after concerns that regulation could hinder innovation.欧盟于2024年通过《人工智能法案》，随后因担心监管可能阻碍创新而推迟了针对高风险系统的规定。
  • On September 29, major AI companies signed a White House accord pledging independent audits and board-level oversight, but the pact carries no penalties.9月29日，主要人工智能公司签署白宫协议，承诺开展独立审计并加强董事会层面的监督，但该协议没有处罚措施。

The three systems differ less in whether AI should be governed than in who sets the rules and how strongly compliance is enforced.三种体系的主要差异不在于是否应监管人工智能，而在于由谁制定规则以及合规要求的执行力度。

023 Dimensions维度分析

Comparison对比

US: voluntary safeguards and executive action美国：自愿保障措施与行政行动

US policy is designed to support rapid AI development through executive orders and industry pledges rather than a comprehensive federal statute. Recent measures address early access to powerful models and independent oversight, but companies face no penalties under the White House accord.美国政策主要通过行政命令和企业承诺支持人工智能快速发展，而不是依靠全面的联邦法律。近期措施涉及提前接触强大模型和独立监督，但白宫协议没有规定企业未履约时的处罚。

Measures in the AI Action Plan《人工智能行动计划》中的措施数量

July 20252025年7月90 plus measures

This year今年30 days

  • Congress has yet to pass a federal AI law, and Trump has pushed to block state AI laws.国会尚未通过联邦人工智能法律，特朗普也推动阻止州级人工智能法律。
  • The September 29 White House accord includes independent audits and board-level oversight but no penalties.9月29日的白宫协议包括独立审计和董事会层面的监督，但没有处罚措施。
Contradiction矛盾之处

China: expanding controls without a general slowdown中国：扩大监管但不普遍放缓发展

China combines detailed controls on content and specific applications with calls for human control, monitoring and emergency systems. Analysts cited in the report say a general pause or deliberate slowdown in AI development remains outside the mainstream policy posture.中国将对内容和具体应用的详细管控，与对人工控制、技术监测和应急系统的要求结合起来。报道引用的分析人士表示，全面暂停或有意放缓人工智能发展，仍不是中国主流政策立场。

  • Rules on AI companion bots aim to curb emotional dependency and prohibit digital clones of real people without consent.针对人工智能陪伴机器人的规定旨在遏制情感依赖，并禁止未经同意创建真实人物的数字克隆体。
  • Xi Jinping has called for AI to remain under human control, alongside stronger laws, monitoring, early warnings and emergency responses.习近平要求人工智能始终处于人类控制之下，并呼吁加强法律、监测、预警和应急响应。
Stakeholder impact利益相关方

EU: rules intended to shape global standards欧盟：以监管规则塑造全球标准

The EU presents its AI Act as a comprehensive framework and is applying additional scrutiny through digital-services rules. At the same time, delays for high-risk-system requirements and concerns from European companies show the tension between stronger safeguards and competitiveness.欧盟将《人工智能法案》定位为全面框架，并通过数字服务规则实施额外审查。同时，高风险系统要求被推迟，以及欧洲企业表达的担忧，都体现出加强保障与维护竞争力之间的张力。

  • The EU adopted its AI Act in 2024 and later delayed rules covering high-risk systems.欧盟于2024年通过《人工智能法案》，随后推迟了针对高风险系统的规定。
  • ChatGPT became the first AI chatbot added to the EU list of digital services subject to greater legal scrutiny.ChatGPT成为首个被列入欧盟数字服务名单、并受到更严格法律审查的人工智能聊天机器人。
  • The EU has proposed requiring AI companions and chatbots to be safe before children use them.欧盟提议要求人工智能陪伴工具和聊天机器人在儿童使用前确保安全。

03Timeline时间线

  1. 2023 The Biden administration issued an order requiring developers of the most powerful AI systems to share safety test results with the government.拜登政府发布命令，要求最强大人工智能系统的开发者向政府提交安全测试结果。
  2. 2024 The European Union adopted its AI Act.欧盟通过《人工智能法案》。
  3. 2025-07 The White House unveiled an AI Action Plan listing more than 90 measures to speed up development.白宫公布《人工智能行动计划》，列出90多项加速发展的措施。
  4. 2026-09-29 Major AI companies signed a White House accord pledging independent audits and board-level oversight.主要人工智能公司签署白宫协议，承诺开展独立审计并加强董事会层面的监督。

Key takeaway核心结论

AI governance is fragmenting into US voluntary commitments, China’s controlled development model and the EU’s rules-based framework.人工智能治理正分化为美国的自愿承诺模式、中国的受控发展模式和欧盟的规则导向框架。

04What to Watch后续关注

  1. Whether Congress passes a federal AI law after the midterm elections.美国国会在中期选举后是否通过联邦人工智能法律。
  2. How state AI laws continue to take effect while Washington seeks to block them.在华盛顿试图阻止州级人工智能法律的情况下，各州法律如何继续生效。
  3. Whether the White House accord produces meaningful oversight despite having no penalties.白宫协议在没有处罚措施的情况下，能否形成实质性监督。
  4. How the EU implements delayed high-risk-system requirements while addressing companies’ innovation concerns.欧盟如何落实已推迟的高风险系统要求，同时回应企业对创新受限的担忧。

05Uncertainties不确定因素

  • The source does not specify when the EU’s delayed high-risk-system rules will take effect.消息来源未说明欧盟推迟的高风险系统规定何时生效。
  • The practical consequences of the voluntary US commitments remain unclear because the accord includes no penalties.由于美国协议没有处罚措施，其自愿承诺将产生何种实际效果仍不明确。
  • The source describes different policy directions but does not establish which approach will better balance safety, innovation and competitiveness.报道描述了不同的政策方向，但没有说明哪种方式能更好地平衡安全、创新与竞争力。

06Sources信息来源

  1. US, China, EU: three paths to regulating AIUS, China, EU: three paths to regulating AI AFP · 2026-10-07

Share this analysis 分享此深度分析

X Facebook LinkedIn WhatsApp