Did You Watermark It? Why That Might Not Be Enough Anymore

The Illusion of Safety: Watermarks as a First Line of Defense

Many platforms, including RedBubble, offer built-in protections for creators. One of the most common tools is watermarking—a visible overlay that signals the work’s ownership and discourages unauthorized use. It’s an easy, accessible safeguard that’s become standard across artist-friendly marketplaces.

But here’s the catch: a watermark can deter only those who care.

In theory, watermarking helps prevent the theft of high-resolution images by making them less appealing to copy. But in practice, it often fails to stand up to infringers armed with basic editing tools.

A Personal Story: When a Watermark Wasn’t Enough

Take my mom’s experience. Her original chinoiserie artwork, shown above, was uploaded to RedBubble with a watermark clearly visible. But that didn’t stop her designs from being ripped off and sold on Temu, stripped of the watermark and repackaged as cheap counterfeits.

Despite RedBubble’s protection settings, infringers found a way around them. Some simply cropped the watermark out. Others used software to blur or erase it. The result? A blatant act of intellectual property theft that not only hurt her as an artist, but also devalued her original work in the eyes of consumers.

Why Watermarks Fail — And What Research Says

Watermarks are one of the first things artists are told to use to protect their work online. They’re free, simple to apply, and send a clear message: This design belongs to someone.

But let’s be real—watermarks don’t actually stop theft.

In theory, they deter casual image stealing. But in reality, even basic editing tools or a bit of cropping can remove or obscure a watermark in seconds. And with the help of AI? It’s even easier.

Research backs this up. A 2017 study by Tali Dekel, Michael Rubinstein, Ce Liu, and William T. Freeman demonstrated that visible watermarks become easy targets when they’re applied consistently across images—machine learning algorithms can learn and remove these watermarks with surprising accuracy[^1]. Even more worrying, a 2025 study by Preston Robinette and Taylor Johnson introduced an advanced AI tool called MorphoMod that can remove watermarks without needing the original image, handling both opaque and transparent marks with over 50% better accuracy than previous methods[^2].

In short: if your watermark looks the same every time—or even if it’s complex—today’s AI tools can erase it, leaving your art exposed.

So why are we still using them? Because for most artists, it’s all we’ve got. Filing DMCA takedown requests takes time. Legal help is expensive. Platform enforcement? Often hit-or-miss. And let’s not forget the elephant in the room: many consumers either don’t notice—or don’t care—if the art they’re buying is stolen.

Watermarks offer a layer of psychological protection. They say, This is mine. But when it comes to actual protection, they’re more of a speed bump than a roadblock.

Moving Forward: The Bigger Problem

My mom’s experience isn’t unique—it’s part of a growing pattern affecting artists everywhere. As counterfeiters become increasingly sophisticated, creators are left constantly playing defense. Until platforms and policymakers step up with stronger, more effective protections, watermarking will remain a limited but necessary tool in an imperfect system.

In the next post, we’ll dive into how often artists around the world face this challenge—and why platforms like Temu continue to profit from these stolen designs with little accountability.

Coming up next:

“When being a Bestseller is not always great”

[^1]: Dekel, T., Rubinstein, M., Liu, C., & Freeman, W. T. (2017). On the Effectiveness of Visible Watermarks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[^2]: Robinette, P. K., & Johnson, T. T. (2025). Blind Visible Watermark Removal with Morphological Dilation.