中国2025年的AI标签规则减少了错误信息传播,但面临逃避工具和执法不连贯的挑战。 China's 2025 AI labeling rules reduced misinformation spread but face challenges from evasion tools and inconsistent enforcement.

中国于2025年9月1日实施了新的AI内容规则, 要求有可见标签和无形元数据, China implemented new AI content rules on September 1, 2025, requiring visible labels and invisible metadata to combat misinformation as generative AI use hit 515 million users. 抖音和快手等平台现在会提示AI来源披露，已有超过1500亿条商品被标记，2.2亿条帖子被标记。 Platforms like Douyin and Kuaishou now prompt AI origin disclosures, with over 150 billion items labeled and 220 million posts flagged. 用户对无标识内容的怀疑率上升近40%, 跨境追踪时间从72小时降至12小时。 User skepticism toward unmarked content rose nearly 40%, and cross-border tracing time dropped from 72 to 12 hours. 然而,日益扩大的灰色市场为去除标签提供了工具,规避战术变得更加复杂。 However, a growing grey market offers tools to remove labels, with evasion tactics becoming more complex. 专家们警告说,标准不一致,处罚不明确,造成监管漏洞,敦促加快标准化和协调执法。 Experts warn inconsistent standards and unclear penalties create regulatory gaps, urging faster standardization and coordinated enforcement.