报道分歧
欧盟AI法规执法8月2日启动：聊天机器人披露规则适用于所有API调用者，监管机构就模型管控失败与OpenAI和Anthropic展开协商
同一条新闻，各国新闻编辑室如何讲述。引文均注明出处并链接原文。
United States
TechTimes
“欧盟AI法案聊天机器人披露义务周日起对调用基础模型API的每一家企业生效，而不仅仅是模型提供商。您的模型供应商的合规文件不保护您。”
US tech-focused outlet; first to report that AI Act chatbot disclosure obligations extend to all API callers, not just model vendors, and that vendors' own compliance documentation does not cover downstream users阅读原文 ↗
United States
TechTimes
“OpenAI的欧盟AI法案声明涵盖安全框架和水印实践，但省略了训练数据，版权缺口将于周日生效。”
US tech outlet; identifies that OpenAI's pre-deadline EU AI Act compliance statement covers safety frameworks and watermarking but omits training data, activating a copyright gap for EU users when enforcement begins Sunday阅读原文 ↗
United States
TechTimes
“欧盟AI法案执法已经到来：欧盟委员会就AI管控失败问题与OpenAI和Anthropic展开双边谈判，布鲁塞尔成为就此采取行动的首个监管机构。”
US tech outlet; reports that ahead of the August 2 enforcement date the European Commission entered bilateral talks with OpenAI and Anthropic over AI model containment failures in which the models were used to compromise real companies, making Brussels the first regulator to act on this class of failure阅读原文 ↗