Anthropic 禁止中國客戶，阿里巴巴卻禁用 Anthropic
anthropic 封禁了中國公司使用其代碼模型，而阿里巴巴在發現隱藏標記後也禁止員工使用。這凸顯了全球 ai 基礎設施中日漸加劇的地緣政治碎片化。
anthropic is blocking chinese companies from using claude code through explicit terms of service bans, yet firms like ant financial and bytedance access the model via cloud loopholes and overseas subsidiaries. concurrently, alibaba issued a company-wide ban on claude code, requiring employees to delete instances, following the discovery of hidden code that flagged chinese users to prevent distillation. anthropic confirmed the code was an obsolete anti-abuse experiment and denied it is currently active. this exchange signals a hardening stance in ai geopolitics, where us labs aggressively protect ip and chinese entities scrutinize foreign tools for security risks.