AI Image Watermarking Faces New Threat From “Unmarker”
IEEE Spectrum - AI
Aug 7, 2025 14:00
Matthew S. Smith
1 views
airesearchieeetechnology
Summary
A new study reveals that while watermarking is widely promoted and mandated as a solution for distinguishing AI-generated images, a universal attack called "UnMarker" can defeat leading watermarking techniques. This finding raises concerns about the effectiveness of watermarking as a safeguard, highlighting the ongoing challenges in reliably identifying AI-generated content and the need for more robust solutions in the AI field.
As AI image generators advance, telling real images from AI-generated images has proven close to impossible. A recent study from Microsoft with 12,500 global participants found that people can detect AI images with an average success rate of 62 percent—not much better than a coin flip. Watermarking is one proposed solution. The European Union’s AI Act mandates watermarking for most AI image generators, and many companies with AI image generators have implemented a watermark or plan to soon do so. Yet this approach might be a dead end, at least according to a paper presented at the 2025 IEEE Symposium on Security and Privacy. It reveals a new universal attack, UnMarker, which defeats leading watermarking techniques. “All the leaders in the field are promoting and investing in [watermarking], with whole teams dedicated to that,” said Andre Kassis, creator of UnMarker and a Ph.D. candidate at the University of Waterloo, in Canada. “Naturally, we want to know, do these systems deliver on t
