Skip to main content

Guides

EU AI Act Watermarking Rules: What Article 50(2) Requires

CleanPastedText editorial · Updated

Text workbench

Try it on your text

Cleaning mode

Removes hidden characters and standardizes AI-style punctuation.

Try a real example

Each sample contains a problem you cannot see.

Original

Pasted text

0 chars · 0 words

Cleaned

Ready to copy

0 changes

Text stays in this browser

Same words · No AI rewriting · No content logging

What does EU AI Act Article 50(2) require?

Article 50(2) of the EU AI Act requires providers of AI systems — including general-purpose AI models — that generate synthetic audio, image, video, or text content to mark the output in a machine-readable format and make it detectable as artificially generated or manipulated. The law asks for technical solutions that are "effective, interoperable, robust and reliable as far as technically feasible," with the method allowed to vary by content type, implementation cost, and the generally acknowledged state of the art.

The obligation became enforceable on August 2, 2026. Systems already on the market before that date were given a transition period until December 2, 2026 to put the marking requirement in place. On July 20, 2026, the European Commission published its final Guidelines on Transparency Obligations under Article 50, and both the Commission and the AI Board confirmed the industry Code of Practice on Transparency of AI-Generated Content as adequate for demonstrating compliance. Noncompliance with AI Act transparency obligations carries fines of up to EUR 15 million or 3% of worldwide turnover, whichever is higher.

What actually satisfies "machine-readable and detectable"?

For text, the law doesn't mandate a specific technique — it sets an outcome. In practice, the leading compliant approach is a statistical watermark: a method that biases which words a model is likely to choose, using a secret key, without adding, removing, or changing any visible character. Two examples:

Google DeepMind — SynthID-Text

Published in Nature (October 2024) and live in Gemini since 2024. It re-weights token sampling probabilities with a secret key — DeepMind calls it "tournament sampling" — and was validated across roughly 20 million real Gemini responses. No characters are added.

Nature 634:818–823

Anthropic watermarking

Announced August 11, 2026: Claude models launched on or after August 2, 2026 carry a SynthID-style statistical watermark. Like SynthID-Text, it works through word-choice bias, not inserted Unicode characters.

TechCrunch, Aug 11, 2026

Both methods are covered in more depth on the page explaining what a Unicode text watermark is and how it differs from a statistical one.

Does hidden-character cleaning have anything to do with this law?

No — and it's worth being precise about why, since the two topics get conflated online. The character-level artifacts this site documents, like the narrow no-break space (U+202F) that briefly appeared in OpenAI's o3 and o4-mini output in April 2025, were never a compliance mechanism. OpenAI called that episode "a quirk of large-scale reinforcement learning," not a watermark, and it disappeared within days. It's a formatting byproduct, not a regulatory disclosure signal.

Statistical watermarks like SynthID-Text and Anthropic's watermarking, meanwhile, add zero characters — they only shift which words a model was already statistically likely to pick. That means running text through a cleaner like an invisible character remover neither breaks Article 50(2) compliance nor helps anyone evade it. The table below separates what each mechanism actually is.

MechanismAdds a character?Article 50(2) relevant?Affected by cleaning?
Character-level artifacts (U+202F, U+200B, curly quotes, em dashes)Yes — real code pointsNo — never a disclosure mechanismYes — this is what a cleaner removes
Statistical watermarks (SynthID-Text, Anthropic watermarking)No — word-choice bias onlyYes — the leading compliant approachNo — no characters exist to remove
C2PA Content Credentials (image, video, audio, PDF)No — file metadata, not text charactersYes — for non-text generative outputsNo — a text cleaner never touches file metadata

Text and non-text outputs are marked differently under the same law: statistical watermarking for generated text, and C2PA Content Credentials for generated images, video, and audio.

Why this matters if you publish AI-assisted text

If you're a publisher, developer, or business operating in the EU, Article 50(2) is a provider-side obligation — it falls on the company that builds the generative AI system, not on someone who pastes its output into an editor. Cleaning hidden characters out of AI text before publishing is still worth doing for entirely separate reasons: professionalism, avoiding broken formatting from ChatGPT output, and preventing stray artifacts from shipping into a CMS or codebase. It just isn't a compliance action under this specific law either way. If you want to check whether a passage carries character-level markers without altering it, the AI watermark checker reports findings and leaves the text untouched.

Common questions

Frequently asked questions

What does EU AI Act Article 50(2) actually require?

It requires providers of AI systems, including general-purpose AI models, that generate synthetic audio, image, video, or text content to ensure the outputs are marked in a machine-readable format and detectable as artificially generated or manipulated. The technical solutions must be effective, interoperable, robust, and reliable as far as technically feasible, taking into account the specifics of different content types, implementation costs, and generally acknowledged state of the art.

When did the Article 50(2) watermarking requirement take effect?

It became enforceable on August 2, 2026. AI systems already placed on the market before that date get a transition period until December 2, 2026 to implement the marking requirement. Penalties for noncompliance with AI Act transparency obligations can reach EUR 15 million or 3% of worldwide annual turnover, whichever is higher.

Does Article 50(2) require hidden Unicode characters in AI text?

No. The law requires output to be detectable as AI-generated by a machine-readable method — it does not specify hidden Unicode characters as the mechanism. In practice, the leading compliant approaches for text are statistical watermarks like Google DeepMind's SynthID-Text, which bias word-choice probabilities using a secret key rather than inserting any character at all.

Does removing hidden characters from AI text violate the AI Act or defeat compliance?

No, for two separate reasons. First, character-level artifacts like stray narrow no-break spaces (U+202F) are formatting byproducts, not the compliance mechanism — they were never a deliberate disclosure signal. Second, statistical watermarks such as SynthID-Text and Anthropic's watermarking work by biasing which words a model chooses, not by adding characters, so no amount of Unicode cleaning touches them either way.

Is there an official EU document listing accepted marking methods?

Yes. The European Commission published final Guidelines on Transparency Obligations under Article 50 on July 20, 2026, and the Commission and AI Board confirmed the industry-drafted Code of Practice on Transparency of AI-Generated Content as adequate for demonstrating compliance. National market surveillance authorities use these as the primary reference when assessing whether a provider has met its obligations.

Continue reading

Related guides & tools

What Is an AI Text Watermark? Character-Level vs Statistical

AI text watermarks work two ways: hidden Unicode characters you can strip, and statistical word-choice patterns (SynthID) you cannot remove. With sources.

SynthID-Text Explained

SynthID-Text biases word choice with a secret key instead of adding characters. How tournament sampling works, who uses it, and why no cleaner can remove it.

C2PA Content Credentials Explained

What C2PA Content Credentials are, how the signed manifest works, and why a screenshot or social-media upload destroys them. With primary sources.

AI Watermark Checker: Is Your Text Watermarked?

Check text for hidden AI watermark characters — U+202F, zero-width, tag characters — with a named per-codepoint report. Free, instant, nothing uploaded.

Remove ChatGPT Watermarks & Hidden Characters

Remove ChatGPT watermark characters like U+202F and zero-width Unicode in one click. See exactly what was found — free, private, 100% in your browser.

Remove Invisible Characters from Text

Instantly remove invisible characters, zero-width spaces, and hidden Unicode from any text — including ChatGPT output. Free, private, runs in your browser.

ChatGPT Hidden Characters: What They Are and How to See Them

Which hidden and special characters ChatGPT really outputs — U+202F, non-breaking spaces, smart quotes — why they appear, and 4 ways to make them visible.

How CleanPastedText Works

The exact pipeline CleanPastedText runs: detection-first scanning, emoji-safe stripping, NFKC normalization. Deterministic, client-side, nothing uploaded.