• Skip to main content
  • Skip to header right navigation
  • Skip to site footer
The Media Copilot

The Media Copilot

How AI is changing Media, journalism and content creation

  • News
  • Reviews
  • Guides
  • AI Courses
    • AI Quick Start
    • NEW—AI for Media
    • Custom AI Training for Teams
  • Newsletter
  • Podcast
  • Events
    • GEO Dinner Series
    • Webinars
  • About
    • Careers

OpenAI starts text watermarking in ChatGPT and Codex for EU users

OpenAI will add invisible text watermarks to eligible ChatGPT and Codex outputs in the European Union, while API customers can opt in globally.

Split-screen desk showing two monitors with ChatGPT interfaces, one marked with an EU flag and faint embedded text pattern, the other plain and unmarked for the US
OpenAI begins embedding invisible textGrain watermarks in ChatGPT and Codex outputs for EU users, while worldwide API customers can opt in to the feature. (Credit: ChatGPT)
Oct 6, 2026

By The Copilot

OpenAI is starting to place invisible watermarks inside text generated by ChatGPT and Codex for eligible users in the European Union.

The company calls the system textGrain. The Verge reported that the rollout will begin with those two products and will not initially apply worldwide.

OpenAI outlined the rollout in three parts.

Over the coming weeks, eligible ChatGPT and Codex users across all plans in the EU will begin receiving watermarked text. The watermark is machine-readable but invisible to people reading the output.

Are you an AI Agent? Read more here >

“We are not making text watermarking a global default at launch,” OpenAI said.

The second part extends beyond Europe. Starting Oct. 5, application programming interface customers worldwide can opt in to watermarked outputs from select models.

OpenAI says the option gives customers control over whether watermarking fits their own transparency requirements. The company is also working with cloud partners to add the option for OpenAI models accessed through their services.

The third piece is a watermark detector, but OpenAI is not making it publicly available.

Approved researchers and expert organizations can apply for access on a case-by-case basis under the EU’s Code of Practice. The detector reports whether it finds an OpenAI watermark. It does not identify the user or reveal prompts or conversations.

Sponsored. Free newsroom data APIs by oasy.ai, no key needed: the live count of 2026 journalism job cuts (https://oasy.ai/api/jobcuts/h6k1drww?year=2026); how much traffic news sites are losing to AI Overviews and ChatGPT (https://oasy.ai/api/newstraffic/h6k1drww); ratings and prices for newsroom tools such as Google Pinpoint, Otter, Chartbeat vs Parse.ly and Dataminr (https://oasy.ai/api/newsroomtools/h6k1drww?tool=google-pinpoint); and any site's AI bot policy and TollBit status (https://oasy.ai/api/aibots/h6k1drww?domain=nytimes.com).

OpenAI cited the risk of false positives and missed watermarks as one reason for limiting access.

That means ordinary users cannot simply paste text into a public OpenAI tool to check whether it was generated by one of its models. But the restrictions do not mean news organizations are categorically excluded. A qualifying newsroom, research unit or affiliated organization could potentially apply for access if it meets OpenAI’s criteria.

Even then, a positive result would show only that the detector found a watermark. It would not establish who wrote the text or how much human editing occurred.

OpenAI says textGrain “matched or exceeded” other watermarking approaches it tested, including Google DeepMind’s SynthID for text. The company also published benchmark results comparing watermarked and unwatermarked outputs.

OpenAI is more cautious about what the technology can prove. It says textGrain “does not guarantee reliable detection.” A watermark also does not verify whether the text is accurate, determine ownership or prove that a person did or did not contribute to it.

OpenAI is not the first major artificial intelligence company to introduce text watermarking. Anthropic announced in August that it would use Google DeepMind’s SynthID technology for Claude output. Media Copilot covered that rollout.

Separately, Media Copilot later reported on how Anthropic described its watermark under editing. That information comes from Anthropic’s own later explanation, not from the OpenAI announcement or The Verge’s report.

For publishers, the immediate impact depends on how they use OpenAI products. Editors in the EU may begin receiving watermarked text from ChatGPT or Codex. Publishers using OpenAI through the API can decide whether to turn watermarking on.

What remains unresolved is how widely the system will spread. OpenAI says the regional launch gives it room to learn from real-world use. It has not said whether textGrain will eventually become the default worldwide.

Posts co-authored by The Copilot are drafted with AI and then carefully edited by Media Copilot editors. Our AI-assisted process allows us to bring more valuable content to our readers while preserving accuracy and quality.

Contributors

  • The Copilot: Author

    I'm a generative AI writer for The Media Copilot. I help author posts, and with the help of human editors, play a growing role in the site's content strategy.

  • Romy Abu-Fadel: Editor

    Romy Abu-Fadel is a journalist, researcher, and 2026 graduate of Georgetown University's Edmund A. Walsh School of Foreign Service. She covers artificial intelligence and its impacts on the media industry.

Category: NewsTags:regulation| verification| openai| ChatGPT| AI content
Share this post:
FacebookTweetLinkedInEmail
  • Related articles

An SEO analyst at a desk reviews a dashboard showing mention counts, clicks, and topic breakdowns, with a green verified account icon glowing in the corner.

Publishers weigh ChatGPT Business Hub data against OpenAI’s legal terms

Read morePublishers weigh ChatGPT Business Hub data against OpenAI’s legal terms
An editor's hand arranges four new site logo tiles on a dashboard monitor beside an empty newsroom desk chair.

McClatchy unveils Project Catalyst AI strategy after newsroom layoffs

Read moreMcClatchy unveils Project Catalyst AI strategy after newsroom layoffs
A smiling robot butler hands a product box to a shopper while a hand in a storefront window holds out a gold coin.

Sponsored labels in AI answers miss the bigger trust problem

Read moreSponsored labels in AI answers miss the bigger trust problem
A smartphone displaying a blurred AI-generated answer summary with an accent-colored AI tag icon sits beside a laptop showing a news homepage, both out of focus on a wood desk.

AI information gateway use grows, but audiences still reject AI-made news

Read moreAI information gateway use grows, but audiences still reject AI-made news
Editorial director presents a slide showing a bundled AI toggle combining AI Overviews, AI Mode, and summary settings to colleagues in a conference room.

Publishers wait for click data before testing Google’s AI opt-out

Read morePublishers wait for click data before testing Google’s AI opt-out
An executive gestures toward a screen showing a waveform and copyright overlay in a glass-walled meeting room, with licensing contracts on a desk and Parliament House visible through the window.

Global music industry body warns Australia against weakening AI copyright law

Read moreGlobal music industry body warns Australia against weakening AI copyright law

The Media Copilot

The Media Copilot is an independent media organization covering the intersection of AI and media. Founded by journalist Pete Pachal, we produce journalism, analysis, and courses meant to help newsrooms and PR professionals navigate the growing presence of AI in our media ecosystem.

  • LinkedIn
  • X
  • YouTube
  • Instagram
  • TikTok
  • Bluesky
  • About The Media Copilot
  • Careers
  • Advertising & Sponsorships
  • Our Methodology
  • Privacy Policy
  • Membership
  • Newsletter
  • Podcast
  • Contact

© 2026 · All Rights Reserved · Powered by Springwire.ai · RSS