OpenAI 说明文本水印工作原理：嵌入统计信号，可检测是否由其模型生成
OpenAI 亲口讲清水印到底能干嘛：只能判断是不是它家模型写的，查不了作者是谁，而且不影响模型速度和文风。
OpenAI 在 X 平台解释其文本水印机制的边界。水印在文本生成时嵌入不可见的统计信号，检测器据此判断文本是否大概率由 OpenAI 模型生成。OpenAI 强调水印不会标注作者身份、归属权，也无法识别个人、组织、账号、对话或提示词。测试显示水印未影响模型能力、速度或回复的可读性，用户在可能的情况下拥有选择权。
When possible, we want to give people a choice—and be clear about what watermarking can and can’t do. A text watermark helps answer one question: “Was this likely generated by an OpenAI model?” It works by embedding an invisible statistical signal in text as it’s generated. A detector can look for this signal, and it doesn’t add visible marks or special characters. It doesn’t say who originally wrote the text or who owns it. It doesn’t identify a person, organization, account, conversation, or prompt. In our testing, watermarking hasn’t affected how capable or fast our models are, or how their responses read. 💬 14 🔄 5 ❤️ 101 👀 18198 📊 20 ⚡
