JD Lasica has a nice pointer to a story about progress in the digital watermarking debate. He wonders about this progress because of work (in part by Ed Felten) suggesting “that all such encryption systems can be defeated.” But there is an important distinction that this debate needs. I’m a strong supporter of flawed (in the sense of defeatable) watermarking. Here’s why:
A watermark is a tag. If it is to work, it needs to be strong enough to resist reasonable incentives to remove it. What incentives are “reasonable” depends upon what the watermarking system is for. (Think of the difference between UPCs on a box of cereal and those chunks of plastic and steel on clothes in a department store.)
The DRM reason for watermarking is to enable digital locks to control access to digital content. The incentives to pick those locks will be strong — especially if prices for content remain high.
But some watermarks enable systems other than locks, and for those, there is very little reason to remove the watermark. These watermark systems in turn can afford to be weak.
For example, William Fisher’s proposal (outlined by Derek Slater here) for compensating artists depends upon a watermarking system. But there would be very little incentive to remove the watermark from the digital objects because the object is free anyway.
Or again, we at CreativeCommons have been presented with lots of useful and clever ideas for marking digital content to enable a simple link to the CC license. As CC licenses are meant to enable the expression of freedoms, not the mechanization of control, there’s again little incentive to remove those tags.
If we had a world filled with weak watermarks (again, like UPCs on products), that would tilt evolution towards systems that depended upon expression rather than control. Some of those systems would benefit compensation regimes (Fisher’s and others); some would enable “some rights reserved” regimes, like Creative Commons. But the more such alternative regimes that we can produce — alternative to the CONTROL model of DRM — the better.
We can have compensation without DRM control. We can have a respect for rights (Creative Commons) without building “the police state into every computer,” as Intel’s Craig Barrett so nicely put it.