OpenAI Is Watermarking ChatGPT Text in the EU and Canada Is Outside the Rule

OpenAI textGrain watermarks ChatGPT text in the EU, not Canada. A missing mark proves nothing. What Canadian firms should do now.
Rubber stamps and import forms on a desk

OpenAI says its textGrain watermark will mark ChatGPT text in the EU, not Canada, and a missing mark proves nothing about who wrote a document. Marcus Laporte reads the fine print for Canadian businesses.

  • Attributed claim: OpenAI says it will add an invisible watermark called textGrain to ChatGPT and Codex text in the EU over the coming weeks, per OpenAI.
  • Verified fact: OpenAI’s own post says API customers anywhere can opt in, and the ChatGPT rollout covers the EU only, so Canadian users are outside it.
  • Analysis: a missing watermark proves nothing, so it cannot be used to accuse a supplier or an employee.

An OpenAI text watermark in the EU is a compliance step and not a Canadian rule, and it will not tell a Canadian employer who wrote a document. OpenAI’s own post on EU text provenance and its provenance help article, which we reread on November 5, 2026, says the ChatGPT rollout is for the EU only and that watermarking will not be switched on worldwide by default. The rule behind it is Article 50 of the EU AI Act, which requires machine-readable marking of AI-generated text and applies from August 2, 2026.

What is textGrain and who gets it?

According to OpenAI, it is applied to ChatGPT and Codex output for EU users over the coming weeks. API developers anywhere can switch it on, but it is off by default. Detection is limited to approved researchers and expert organizations, so an ordinary company cannot check a document.

Advertisement

Key points

  • OpenAI is rolling textGrain watermarks onto eligible ChatGPT and Codex text in the EU only; Canada is outside that ChatGPT rollout.
  • API watermarking is opt-in worldwide and off by default.
  • A missing watermark does not prove a person wrote the text, and OpenAI says light synonym edits can drop detection sharply.
  • Canadian firms should write their own disclosure rule and ask vendors whether API watermarking is on, not treat detectors as proof.

Canadian ChatGPT users are outside the EU rollout, so their text is not marked under this programme.

Does this change anything for a Canadian company?

Directly, no. Canada has no equivalent duty, and our earlier look at Bill C-36 explains why federal AI rules remain thin. Indirectly, yes in two cases. If you sell into the EU, you may be asked how your AI generated content is labelled.

And if a supplier builds on the API, ask whether it switched marking on, since the default is off.

Why an unmarked document proves nothing

The limits are the story. OpenAI states that the lack of a watermark is no proof that a person wrote the text, because short passages, maths answers, translations, edited text and other vendors’ models all weaken or avoid the signal. That cuts both ways. A marked document proves an OpenAI model was involved somewhere, and an unmarked one proves nothing. Anyone building a policy that treats detection as proof will punish the wrong people.

What should you do this month?

Write your own one-page rule on disclosing AI use to customers, based on your contracts rather than on a detector. Then ask vendors whether watermarking is switched on, check any EU transparency duties with counsel, and take detector results out of HR and grading decisions. This month’s checklist:

  1. Write a one-page rule on disclosing AI use to customers, based on your own contracts, not on a detector.
  2. Ask any vendor using the OpenAI API whether watermarking is on and who can verify it.
  3. If you sell in the EU, ask counsel which transparency duties reach your content and from what date.
  4. Remove detector results from any HR or grading procedure.

What does the sceptic say?

The sceptic says watermarks are a first step toward trust in AI content, and that Canada will copy the EU. Both may be right over time. But OpenAI’s own evaluation says replacing about 10% of words with synonyms cut detection from about 92% to 66% on 400-token passages, so a first step like that is not yet workplace evidence, and no Canadian bill proposes it.

What would prove me wrong is independent testing showing high detection after realistic edits.

What to watch

Watch for the first independent tests of textGrain, whether other vendors follow, and whether Ottawa mentions marking in a future AI bill. Also watch the EU timetable, since reports differ on when each obligation applies.

Who in Canada should care first

The Office of the Privacy Commissioner’s 2024-25 annual report says 83% of Canadians have some concern about their privacy when using AI tools. Readers who worry about AI will ask whether a text was machine-made, and a watermark that stops at the EU border gives them no answer here.

Businesses that exported services outside Canada used AI at 40.9%, more than double the 19.2% average, according to Statistics Canada table 33-10-1167-01. Those are the firms most likely to send ChatGPT-drafted text to European clients, so they are the ones to tell staff what textGrain is.

Frequently asked questions

Does ChatGPT watermark text in Canada?

No, not under this programme. OpenAI says the textGrain ChatGPT rollout covers the EU, and Canadian users are not included.

Can I tell if a document was written by ChatGPT?

No, not reliably. OpenAI says the lack of a watermark is no proof that a person wrote the text, and detection is limited to approved researchers.

Does the EU AI Act apply to Canadian businesses?

It can, if you place AI systems or their output on the EU market. Ask counsel which duties reach your case and from what date.

The decision in one line

Write your own AI disclosure rule now, and never treat a watermark or its absence as proof of who wrote something.

Archive entry dated October 6, 2026, written and fact-checked on October 7, 2026. Sources are linked on the claims they support.

Total
0
Shares
Prev
ChatGPT Image Ads Are a Test Your Brand Should Watch, Not Buy
Blank glowing billboard at a wet bus stop at night.

ChatGPT Image Ads Are a Test Your Brand Should Watch, Not Buy

OpenAI will test visual ads beside ChatGPT image results in the US later in

Next
Mistral Large 4 Gives Canadian Teams a European Open-Weight Option With Fine Print
Wind turbines in a green field under blue sky.

Mistral Large 4 Gives Canadian Teams a European Open-Weight Option With Fine Print

Mistral Large 4 launched in public preview on October 6 with a 1 million token

You May Also Like