OpenAI explains what its text watermark can and can't do

The watermark embeds an invisible statistical signal as text is generated and answers only whether text was 'likely generated by an OpenAI model'; it does not identify a person, account, conversation or prompt, and in testing did not affect capability, speed or style. A follow-up post (~173k views) says watermarks are often undetectable in short passages and can be removed by rewriting or translation.

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.

views 225828 · likes 634 · reposts 36 · replies 53 (at fetch time)

Archived 2026-10-06 via fxtwitter (unofficial).