OpenAI Watermarks ChatGPT Text, Expands Ads: What UK Marketers Need To Know

Right, OpenAI just dropped two pretty big bits of news that UK marketers need to pay attention to. On 6 October 2026, they announced they're rolling out invisible watermarks for ChatGPT text.

This isn't just a technical tweak; it's about compliance for any content you're creating for European audiences and new opportunities to reach customers. If you're using AI for content or thinking about where your ads could appear next, you'll want to take note.

]What's actually changed

OpenAI is introducing invisible watermarks to text generated by ChatGPT and Codex. This change is mandatory for users in the European Union (EU) and will roll out “over the coming weeks” to comply with the EU AI Act. OpenAI tweeted that in the EU, they'll “start watermarking eligible text from ChatGPT and Codex over the coming weeks to comply with the EU AI Act.”

The watermarking system, which OpenAI calls “textGrain”, subtly influences the word choices the AI makes. This creates a statistical pattern that a detector can later pick up, even if you can't see it. As of 5 October 2026, API customers worldwide can opt-in to watermarking for selected models, though it remains off by default.

Now, about detection: OpenAI's own tests show their detector identified watermarks in approximately 80% of 200-token passages and 95% of 400-token passages in psychology-related content, at a 1% false-positive rate. However, editing AI-generated text can significantly reduce detection. OpenAI found that replacing 25% of words in a 400-word passage reduced detection to just 17%.

OpenAI also stated, quite plainly, that “The absence of a detected watermark does not prove human authorship.” They also clarified that the watermark “does not indicate who originally wrote the text or who owns it,” and “does not identify a person, organisation, account, conversation, or message.” They've also been upfront, noting that “Text watermarking and detection remain early technologies with significant limitations.” Access to the watermark detector will be limited to approved researchers and expert organisations, with applications now open.

Why this matters to UK marketers

Even though the UK isn't part of the EU, many UK businesses target European audiences. If you're creating content with AI for those markets, you need to be aware of the watermarking. The EU AI Act is a big piece of legislation, and its reach can extend beyond EU borders for services consumed there.

This means if your AI-generated content is going out to EU customers, it could fall under these new transparency requirements. The fact that the watermark is invisible, and that its detection can be weakened by editing, makes it a bit of a grey area. But the direction of travel is clear: transparency for AI-generated content is coming.

The bigger picture for AI content

What we're seeing here is just the beginning of how AI-generated content will be treated and regulated. Transparency, attribution, and monetisation are all going to become bigger conversations.

For content creators, the human touch, editing, refining, and adding unique insights, remains vital, not just for quality but for understanding these new technical and regulatory challenges. For advertisers, AI platforms are evolving into powerful new media channels, offering fresh ways to connect with customers.

This is all moving quickly, and staying on top of these changes is key. We can help you manage the complexities of AI in your marketing strategy.