PolicyTechCrunch

Anthropic shares more details about how Claude’s new watermarks will work

#ai#watermarking#anthropic#claude#eu ai act

English

Anthropic has detailed the watermarking system for its AI chatbot Claude, which aims to comply with the EU AI Act’s Transparency Code. The watermark will be undetectable to readers but identifiable to those with a key, and while light editing may not remove it completely, a full rewrite would. The company plans to release a watermark detection API and noted that other AI developers will also implement similar watermarks.

中文

Anthropic 详细介绍了其 AI 聊天机器人 Claude 的水印系统，旨在遵守欧盟 AI 法案的透明度规范。水印对读者不可检测，但可以被拥有密钥的人识别，轻微编辑可能无法完全去除水印，而完全重写则会去除。该公司计划发布水印检测 API，并指出其他 AI 开发者也将实施类似水印。