On August 14^th, Anthropic announced that they would be implementing a watermark for content created or heavily edited by Claude to abide by the EU Artificial Intelligence Act. For those on your team who’ve been skeptical of AI from the start, we’re sure that’s created even more confusion around when and how we should use AI tools in content creation. To help clear things up, GroundFloor Media will walk you through what the Claude watermark actually does and what this means for you.  

The Claude Watermark At a Glance 

The Pew Research Center cites that about 35% of webpages in its sample that were published after the release of ChatGPT in 2022 show clear signs of AI authorship, which is a statistic consistent with the findings of other studies. With a change in the way content is influenced by AI on the internet, it’s crucial for your brand to ensure you’re sending out content that still reflects the people, perspective and expertise behind your work.   

The way that the Claude AI watermark works is by writing one word at a time, then several words that would work as the next option and chooses at random. For example, think of all the different ways you could describe a rainy morning. Now, this randomness will be replaced with a hidden pattern that indicates to tools that AI was involved in its creation.  

How to find the Claude Watermark 

The Claude watermark is not something visible to the human eye that appears on content, rather a way that content is flagged with a code if it’s been heavily edited by AI. That being said, creative or open-ended content is likely to be more susceptible to watermarking because the language processing model must decide on more word choices. Highly factual content, direct quotes or work where Claude was only cleaning up errors leaves very little trace. 

What AI watermarking means for you 

The good news is that this shouldn’t change what you’re doing if you’ve already been thoughtful about how AI is integrated into your work. At GroundFloor Media, we’ve always believed AI belongs in the workflow as a partner and a tool but should not substitute as a content creator. Research and ideation are excellent examples of how an organization might use AI to help improve the structure and efficiency of projects. When words, ideas and voice start coming from a robot, we start to lose the human touch that’s so important to building connections. That’s when the watermark catches it.  

Just because Claude is the first provider to announce this change doesn’t mean they be the only one. The EU Act that drove the new watermark applies to all companies, and each of them will have to have a solution soon. Even without the watermark increasing AI detection, without heavy human input, your content would just end up looking like everyone else’s, muting the impact that it can have in the first place.  

What you should do now 

For most organizations, this update will have less of an impact than you might think. You can think of if it as a safeguard to ensure your team’s work is fundamentally your own, with the ability to keep using Claude as a tool to improve efficiency. In an oversaturated media landscape, real content made by real people is what your audience will remember.  

If you’re curious about what tools you could use to streamline your content strategy, we’d love to have that conversation with you.