I feel like the article is focusing on the wrong target here: the watermarks are not the problem. The problem is the vast and interconnected ubiquitous surveillance machine that makes it possible to track these watermarks around and, therefore, determine who is talking to whom about what.

While we can’t make use of the information in the mark ourselves, it seems like, if we can figure out how to spot them, it should be possible to make watermarks with fake but legitimate-seeming values in other to confuse things.

Spymarks, brought to you by society’s skidmarks.

That’s some scary stuff, for people that need to share things in secret. Guess the best way is to take photos on an older camera, but for documents with hundreds of pages, that definitely not easy

Great read. I wonder if there’s a possibly for open source tools to decode these types of marks? I can’t imagine it’s easy especially if a company like google is creating it so that they have proprietary access to the stored information