Anthropic 正在推出其水印验证 API，让已批准的组织可以检查文本中是否包含 Claude 的数字水印。自 2025 年 8 月 2 日起，《欧盟人工智能法》要求新款 Claude 模型在其文本输出中嵌入不可见水印：https://the-decoder.com/anthropic-watermarks-all-claude-outputs-globally-with-marks-that-may-persist-through-some-editing/。"监管机构、执法部门、媒体、事实核查人员、独立研究人员、教育机构以及欧盟民间社会团体"都可以申请访问：https://forms.gle/9tGA33hPJJwtHsMk9，企业若需验证水印以确保自身合规，也可申请访问。Anthropic 计划在未来逐步扩大访问范围。
该系统基于谷歌的 SynthID 文本方法：https://the-decoder.com/google-deepmind-open-sources-ai-text-watermarking-for-gemini/，通过调整词汇选择的随机性创建一个可统计检测的模式。据 Anthropic 称，这种模式"可能在某些编辑后依然存在"。这让它比传统检测器如 Pangram 更可靠：https://the-decoder.com/pangram-says-its-new-ai-text-detector-makes-only-one-mistake-per-24000-documents/。
Anthropic：https://www.anthropic.com/news/claude-text-watermark 表示水印不包含用户数据，并且不会影响质量或内容。批评者持不同意见：https://the-decoder.de/warum-claudes-unsichtbare-wasserzeichen-weder-unsichtbar-noch-harmlos-sein-koennten/：如果 Claude 根据水印密钥而非意思选择同义词，文本质量将受影响。行业出版物《Artificial Lawyer》也指出透明性风险，因为可检测的 AI 指纹可能在合同禁止使用 AI 或费用谈判时成为问题，不过预计水印大多无害。广告
保持对 AI 的了解。内容清晰、有用，无冗余。
关注 The Decoder 获取 AI 新闻、背景故事和专家分析。
解码器：https://the-decoder.com/
Anthropic is launching its watermark verification API, letting approved organizations check whether text contains a digital watermark from Claude. The EU AI Act since August 2, 2025, requires new Claude models to embed invisible watermarks in their text output：https://the-decoder.com/anthropic-watermarks-all-claude-outputs-globally-with-marks-that-may-persist-through-some-editing/. "Regulators, law enforcement, media, fact-checkers, independent researchers, educational organizations, and EU civil society groups" can request access：https://forms.gle/9tGA33hPJJwtHsMk9, as can enterprises that need to verify watermarking for their own compliance. Anthropic plans to expand access over time.
The system builds on Google's SynthID text method：https://the-decoder.com/google-deepmind-open-sources-ai-text-watermarking-for-gemini/, tweaking word-selection randomness to create a statistically detectable pattern that, according to Anthropic, "may persist through some editing." That should make it far more reliable than traditional detectors like Pangram：https://the-decoder.com/pangram-says-its-new-ai-text-detector-makes-only-one-mistake-per-24000-documents/.
Anthropic：https://www.anthropic.com/news/claude-text-watermark says the watermark contains no user data and affects neither quality nor content. Critics disagree：https://the-decoder.de/warum-claudes-unsichtbare-wasserzeichen-weder-unsichtbar-noch-harmlos-sein-koennten/: if Claude picks synonyms based on a watermark key rather than meaning, text quality suffers. Trade publication "Artificial Lawyer" flags transparency risks too, since detectable AI fingerprints could become a problem where contracts ban AI use or during fee negotiations, though it predicts the watermarks are mostly harmless. Ad Ad
Stay in the loop on AI. Clear, useful, no fluff.
Follow The Decoder for AI news, background stories and expert analyses.
The Decoder：https://the-decoder.com/
扫码直达文章详情