# A short countdown to Article 50 text marking

Originally published at https://gifi.ai/blog/article-50-text-marking-countdown

Between August and 2 December 2026 the useful work is naming which lab marks which surface — not buying a generic “AI watermark remover.”

## The next hundred days

Providers who already mark text will keep doing it. Providers who do not will have to say how they will. Users will keep searching for a single remover that works on every lab. That search is the confusion.

## A better question

Ask which surface you are looking at. Words, file metadata, or pixels. Then ask which lab. Claude and Gemini watermark text. ChatGPT does not. A detector score is not a watermark.

## What we are publishing

The cluster lives under /guides. The dated notes stay here. If you arrived from a headline about 2 December, start at /guides/eu-ai-act-text-marking and /guides/does-chatgpt-watermark-text.

## Primary sources

- [EU Code of Practice signatories](https://digital-strategy.ec.europa.eu/en/news/strong-backing-code-practice-transparency-ai-generated-content)
- [Anthropic on Claude text watermarking](https://www.anthropic.com/news/claude-text-watermark)