
What happened
OpenAI said it will add an invisible watermark to ChatGPT and Codex text in the EU, rolling out over the coming weeks, and published a technical report for its method, textGrain.
Why it matters
OpenAI also said the watermark does not identify the user and that it saw no meaningful change in model performance with it on, reducing concerns about the trade-off.
What to watch
OpenAI's tests suggest editing can weaken the mark, with synonym swaps dropping detection from about 92% to 66%. Watch whether that limits the watermark's reliability in practice.
WHO IT HITSThis affects EU users of ChatGPT and Codex on all plans, plus developers worldwide using OpenAI's API who can now turn watermarking on for select models. It also matters to researchers and expert organizations approved for detector access.
Summaries like this, every morning.
OpenAI's decision to watermark text in the EU follows the EU AI Act's transparency rules, which took effect on August 2. The company is rolling this out only in the EU, while API developers worldwide can enable it for select models starting today. Two months ago, Anthropic said it would watermark Claude text worldwide, drawing backlash from some users. OpenAI had built a text watermark before but held off, partly over concerns users would switch to rivals that didn't watermark, according to The Wall Street Journal in 2024. The watermark works by subtly shaping word choices, so it travels with copied text. OpenAI's tests show editing can weaken detection, which may limit reliability. The outcome hinges on whether the watermark remains detectable in real-world use and whether approved researchers can validate it.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Altimeter's Brad Gerstner said applying a 10x multiple to 2028 numbers gets Anthropic to about $3 trillion in…

Dragos sent roughly 2 million AI requests with fake credentials from July through November 2026 and found at l…

Following Jensen Huang's ouster as Nvidia CEO, investor Bill Birmingham of REX Financial said the successor in…

SpaceX announced on Nov

The author lays out the exam mapping — Trainium (Trn series) for training, Inferentia (Inf series) for inferen…

Wells Fargo's updated terms, effective Tuesday, put the risk of third-party AI agents on customers, after JPMo…
