Anthropic has begun building a watermark into text generated by future Claude models, a change the company says is meant to help identify whether a given piece of writing was likely produced by its AI. This new feature, implemented to comply with EU rules, is meant to be indistinguishable to the human eye, without changing Claude's normal writing output.
The company laid out the mechanics and rationale behind the feature in a post published to its website.
How does the watermark work?
According to Anthropic, the watermark exploits the countless small, low-stakes decisions a language model makes as it generates text. Rather than using a truly arbitrary random number to make that pick, the watermarked version of Claude bases the decision on a cryptographic key combined with the preceding text.
The result, Anthropic says, is a subtle statistical pattern spread across a response that's invisible to a human reader but detectable to anyone with the matching key, which allows them to estimate the probability that Claude generated the text.
Does it cost more or slow Claude down?
Anthropic was clear in that the change carries no cost to output quality. The company said internal testing turned up no measurable difference in the creativity, accuracy, or readability of watermarked versus unwatermarked responses, and pointed to findings from Google DeepMind's original research on the underlying technique — the method Claude's watermark is based on.
The company also said that its researched showed no statistically significant shift in user satisfaction when a similar watermark was tested on live traffic. Anthropic also said the feature adds no extra tokens, meaning it doesn't slow Claude down or make it more expensive to use.
Where does the watermark break down?
Like with all tools, the watermark has limits. Anthropic explained that it only works when a model is choosing among several equally valid options, so text with little room for variation, such as hard factual statements, precise code, or math answers, carries a much weaker or nonexistent signal.
Detection also grows less reliable on very short passages, since there's simply less pattern to analyze. So you'll get a much clearer read on Claude's likely involvement in longer-form text. And because the watermark tracks only the words Claude itself selects, lightly edited or proofread human writing may carry little to no detectable trace, since most of the original wording remains unchanged.
A sufficiently heavy rewrite, the company noted, can remove the watermark entirely. At that point, per Anthropic, it becomes debatable whether the resulting text is still meaningfully AI-generated.
Can the watermark identify my organization or me?
Anthropic stressed that the watermark can't be traced back to a specific user, account, or conversation, and that it doesn't establish authorship or ownership over content. All it can tell you is the likelihood that Claude was involved in producing or editing it at some point.
The company also distinguished the approach from third-party AI-detection tools, which typically rely on spotting stylistic patterns in AI writing rather than checking for an embedded signal tied to a private key.
Why is Anthropic doing this?
The rollout is tied to regulation rather than a purely voluntary move: Anthropic said it signed the European Union's Code of Practice on Transparency of AI-Generated Content in July 2026 alongside roughly 190 other signatories, following an EU AI Act requirement, effective Aug. 2, that AI providers mark generated text.
Because the company doesn't yet have a reliable way to apply the watermark only within the EU, Anthropic said it's rolling out the feature globally and plans to extend it to older Claude models over the coming months.
The company also said it will soon offer a separate API allowing anyone to check whether a piece of text carries Claude's watermark.
Want to learn more about getting the best out of your tech? Sign up for Mashable's Top Stories and Deals newsletters or get Mashable push alerts today.