Earlier this month, Anthropic started placing an imperceptible label on all text produced using its Claude AI model to serve as an indicator that the material was generated with AI. The folks who don\u2019t want it outed that they are using AI for work are now spending their hours looking for a way to scrub their outputs of that marker.<\/p>\n
As Business Insider notes, the term \u201cAI watermark remover\u201d has started trending on Google<\/a>, jumping 60% in search interest week over week, and an open-source GitHub project called Watermarks Remover<\/a> has amassed more than 15,000 \u201cstars\u201d and thousands of forks with the promise to scrub any output of indicators that it came from AI. Sabrina Ramonov, an AI educator, created her own web-based tool<\/a> for washing text of residual model markers.<\/p>\n Why the panic, exactly? Well, according to the creator of the Watermarks Remover tool, it\u2019s a matter of privacy and attribution. \u201cI am all for content attribution,\u201d Guillaume Meyer, the creator of the popular GitHub project, told Business Insider. \u201cI am against the watermarking technique, and that\u2019s a very significant distinction.\u201d<\/p>\n It\u2019s probably fair to say that an AI watermark will create a false binary in many people\u2019s minds. If some content carries the invisible marker, they\u2019ll simply assume the entirety of that material was created with Claude, when it\u2019s possible that only a small portion of it was AI-generated. And it\u2019s not like people aren\u2019t already out here trying to hit people with AI accusations. One only has to look at how popular Pangram and other so-called \u201cAI detection\u201d tools have gotten to see how weaponized something like this could be (albeit, most of the time the stakes are pretty low and amount to little more than \u201cAI shaming\u201d).<\/p>\n Likewise, there is concern that by slapping this marker on content, Anthropic and other companies that institute a similar watermark may be laying the groundwork for claiming ownership over certain material. Anthropic seems to have thought about that, and mentioned in a blog post<\/a> that \u201cA watermark only helps test whether Claude might have produced or processed the content. It doesn\u2019t say anything about ownership or authorship and doesn\u2019t change a user\u2019s rights under our terms.\u201d Of course, that\u2019s only true until it isn\u2019t. A blog post isn\u2019t binding.<\/p>\n