> ## Content Index
> Fetch the complete content index at: https://www.promptthemarket.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# OpenAI watermarks ChatGPT text in the EU, but editing erases it
- URL: https://www.promptthemarket.com/openai-watermarks-chatgpt-text-eu-editing-erases-it/
- Published: 2026-10-12T03:00:00.000Z
- Updated: 2026-10-12T03:01:06.000Z
- Description: Australian ChatGPT users aren't covered, but API customers anywhere can opt in. OpenAI's own tests show a lightly edited draft loses most of the mark.
- Author: Editor News
- Tags: AI, Europe, OpenAI, #author-xaviery-malinao

By **Xaviery Malinao** · 2026-10-11

TLDR

OpenAI said on 5 October it will add textGrain, an invisible statistical watermark, to ChatGPT and Codex output for eligible EU users over the coming weeks, to meet Article 50 of the EU AI Act. Editing just a quarter of words drops detection from 95% to 17%, limiting its use in professionally revised copy. Australian users aren't covered, though API customers anywhere can opt in.

KEY TAKEAWAYS

01textGrain detects watermarks in about 95% of unedited 400-token passages, falling to 17% after a quarter of words are replaced.

02EU AI Act penalties for breaching Article 50 transparency rules reach €15 million or 3% of global annual turnover.

03API customers worldwide can opt into watermarking from 5 October 2026, though it remains off by default.

04Detector access is restricted to approved researchers; no public tool exists to check a passage.

05Australian ChatGPT users aren't covered; API customers worldwide can opt in.

OpenAI will hide a statistical watermark in text written by ChatGPT and Codex for users in the European Union, but its own tests show the mark fades fast once someone edits the copy.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)

The company said on 5 October that the watermark, called textGrain, would be added to eligible output "over the coming weeks" for EU users on all plans. It is meant to meet Article 50 of the EU AI Act, whose breaches can draw fines of up to €15 million or 3% of global turnover.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)[\[2\]](https://cybernews.com/news/chatgpt-watermark-openai/?ref=promptthemarket.com)

"Our text watermarking technology, textGrain, adds an invisible statistical signal to the model's word choices," OpenAI said.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)

## A quarter of the words, and it's mostly gone

At a 1% target false-positive rate, the detector found the mark in about 80% of 200-token passages and about 95% of 400-token passages.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)

Editing changes that. In 400-token passages, swapping 10% of words for synonyms cut detection from about 92% to 66%, OpenAI said. Replacing a quarter of the words cut it to 17%.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)

OpenAI was blunt about the limits. "A watermark does not measure human contribution," it said, and "the absence of a detected watermark does not prove human authorship."[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)

Nobody outside a small group can run the check yet. Access to the detector "will initially be limited to approved researchers and expert organizations", and OpenAI said it plans to release the technology as open source later. A technical report sets out how it works.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)[\[3\]](https://cdn.openai.com/pdf/e9508624-d767-41b6-a26d-e34ca798ada6/textgrain-entropy-calibrated-watermarking-for-language-model-text.pdf?ref=promptthemarket.com)

textGrain detection rate on 400-token passages

Unedited95%

10% of words swapped66%

25% of words replaced17%

At a 1% target false-positive rate. The 10% test started from a baseline of about 92%. Source: OpenAI, 5 October 2026.

## Optional outside Europe

Australian ChatGPT users aren't covered by the rollout. API customers anywhere can opt in to watermarking for selected models from 5 October, but it stays off by default.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)

That leaves Australian agencies that write for European clients, or build products on the API, with a choice to make about the setting. For anyone else, OpenAI's figures cut both ways: a clean detector result proves nothing about who wrote a piece, and a lightly edited AI draft will often pass as unmarked.[\[1\]](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)

SOURCES & CITATIONS

1. [OpenAI: EU text provenance (5 October 2026)](https://openai.com/index/eu-text-provenance/?ref=promptthemarket.com)
2. [Cybernews: OpenAI to watermark ChatGPT text in the EU (7 October 2026)](https://cybernews.com/news/chatgpt-watermark-openai/?ref=promptthemarket.com)
3. [OpenAI: textGrain technical report (PDF)](https://cdn.openai.com/pdf/e9508624-d767-41b6-a26d-e34ca798ada6/textgrain-entropy-calibrated-watermarking-for-language-model-text.pdf?ref=promptthemarket.com)

FREQUENTLY ASKED QUESTIONS

What is textGrain and why has OpenAI released it?

textGrain is an invisible statistical watermark that OpenAI embeds in the word choices of ChatGPT and Codex output. OpenAI released it in the EU to meet the transparency obligations of Article 50 of the EU AI Act (Regulation (EU) 2024/1689).

How accurate is the textGrain detector?

At a 1% false-positive rate, the detector identifies about 95% of unedited 400-token passages. That rate falls to roughly 17% once a quarter of the words in a passage have been replaced, according to OpenAI's own tests.

Can Australian agencies or brands access the detector?

No. Access to the detector is initially limited to approved researchers and expert organisations. No public tool is available. Australian users are also excluded from the mandatory watermarking rollout, though API customers worldwide can opt in from 5 October 2026.

[Xaviery Malinao](https://www.promptthemarket.com/xaviery-malinao/)

Xaviery Malinao writes for Prompt the Market on how brands and agencies are adapting to answer engines, drawing on Bushnote's work with clients across search, AI search and content.