纽约时报报导，中国对人工智能存在性威胁论持怀疑态度，认为AI是可控的发展工具而非末日性风险。1中国对美国AI企业领袖发出的安全警告持警惕态度，认为这些警告意图遏制中国的AI发展。1Anthropic执行长阿莫戴等多名AI巨头曾呼吁限制AI发展速度，1阿莫戴更在英国金融时报撰文提议保持民主国家相对于威权国家的AI领先优势，包括禁止向中国出售强大AI芯片。1
为应对AI安全问题，中国已更新了"人工智慧安全治理框架"，明确要求标识AI生成内容，并要求模型公开发布前必须通过官方备案与审批。1然而报导指出，中国在AI透明度和强制安全测试方面仍存不足。1据安远AI统计，2025年6月至2026年5月期间，中国前十大人工智慧企业中仅有五家公开相关评估结果。1此外，中国尚未对AI系统被用于制造化学或生物武器等灾难性风险设定强制测试要求。1
China maintains skepticism toward artificial intelligence doomsday scenarios, regarding AI as a controllable development tool rather than an existential threat.1 The country views safety warnings issued by American AI company leaders as a pretext to constrain China's technological advancement.1
In response to these concerns, China has updated its "Artificial Intelligence Safety Governance Framework," requiring that AI-generated content be clearly labeled and that models undergo official registration and approval before public release.1 However, the framework still falls short in areas of transparency and mandatory safety testing.1 According to data compiled by Anysis AI, only 5 out of China's top 10 artificial intelligence enterprises publicly released relevant assessment results between June 2025 and May 2026.1 Notably, China has not yet implemented mandatory testing requirements for catastrophic risks such as AI systems being weaponized to produce chemical or biological weapons.1
Dario Amodei, CEO of Anthropic, has argued in an article published in the Financial Times that democratic nations should maintain their AI leadership advantage over authoritarian countries, including restrictions on selling powerful AI chips to China.1
评论
还没有评论，欢迎留下第一条。