How Claude's Text Watermarking Works
Anthropic explains the implementation of text watermarking for Claude, a technique that embeds an imperceptible pattern in generated text to help identify AI-generated content, complying with the EU AI Act. The watermark doesn't affect output quality, cost, or speed, and can't be traced to individual users. It leverages low-stakes word choices to create a detectable pattern, but has limitations with short texts, factual passages, and code.
Watermarking uses low-stakes choices like these—which occur many times over a piece of generated text—to leave a pattern in Claude’s responses.
- absoluteunit1
> Google DeepMind tested this impact by serving a model that used watermarking to a portion of their Gemini traffic and comparing thumbs-up and thumbs-down ratings. They found no statistically significant differences from the unwatermarked model. And in a controlled study, human raters comparing watermarked and unwatermarked answers side-by-side saw no difference in quality.
For some reason I had assumed testing this would be more sophisticated than just checking the thumbs up/down stats and user "vibes"
- pr337h4m
We are very fortunate open source models have reached parity for virtually all non-coding use cases.
- arjie
Interesting. Here's the section of the EU Act that mandates this:
> Providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content, shall ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated. Providers shall ensure their technical solutions are effective, interoperable, robust and reliable as far as this is technically feasible, taking into account the specificities and limitations of various types of content, the costs of implementation and the generally acknowledged state of the art, as may be reflected in relevant technical standards. This obligation shall not apply to the extent the AI systems perform an assistive function for standard editing or do not substantially alter the input data provided by the deployer or the semantics thereof, or where authorised by law to detect, prevent, investigate or prosecute criminal offences.
https://eur-lex.europa.eu/eli/reg/2024/1689/2026-07-27/eng
It definitely makes Pangram's job a bit easier.
- jluysvi
Opus 5 must be the pilot becuase it's writing style is so grating it has to be intentional. Let's hope they make it more subtle in the future.
- WaitWaitWha
Is there a method to opt out of this for non-EU people? I do not see such option in the article.