How Google's ContentID System Fails At Fair Use & The Public Domain
from the it's-a-problem dept
We recently covered how YouTube briefly pulled down the NASA-uploaded public domain video footage from the Curiosity rover’s Mars landing. We were quite careful in the piece not to call it a DMCA takedown, because it was pretty clear that the DMCA was not involved. Unfortunately, many have been assuming that it was via the DMCA (and there are even lengthy comments discussing aspects of the DMCA). However, the DMCA had nothing to do with it. It appears that the whole thing was due to the way that YouTube’s ContentID system works.
Tim Lee has a great post explaining how ContentID works in such situations, including the story of another video — which involved commentary on the Curiosity landing done by Lon Seidman of the site CT Tech Junkie, which quickly received five claims from media organizations to copyright in the content.
Content ID, by contrast, is an opaque and proprietary system where the accuser can serve as the judge, jury, and executioner. Worse, the person whose speech is being silenced has little recourse. The Content ID system tips whatever balance is present in the DMCA and allows even more pernicious forms of manipulation and abuse. In a Wired column earlier this year, Andy Baio enumerated some of the problems that YouTube users encounter:
But there has been a dramatic rise in Content ID abuse in the past couple of years, wielded in ways never intended. Scammers are using Content ID to steal ad revenue from YouTube video creators en masse, with some companies claiming content they don’t own deliberately or not. The inability to understand context and parody regularly leads to “fair use” videos getting blocked, muted or monetized.
But even without taking scammers into account, the premise behind Content ID is just incompatible with fair use and the public domain. It’s impossibly complicated to define in a set of “business rules” for automated enforcement. Allowing Content ID robots to apply the rules leads to oversimplification that chills legitimate speech.
If anything, as Tim Lee’s article explains, ContentID is actually demonstrating (quite clearly) why there are so many concerns about copyright takedowns. Copyright system supporters often insist that it’s “easy” for sites to recognize and take down infringing content, and use any evidence of infringement as a damning sign of a site not doing enough. But, the reality on the ground is that making a determination on whether or not something is infringing is not nearly as easy as some people believe:
But in accommodating the demands of large copyright holders, YouTube has inadvertently reminded us all of the crucial point that flagging copyright infringement isn’t nearly as simple as it is often portrayed by rightsholders. Even scanning videos for exact content matches that exceed certain thresholds (in order to preserve at least some fair uses) actually fails in all sorts of interesting ways.
Rather than acting as a neutral arbitrator between major content companies and independent organizations, YouTube’s system favors the larger rightsholders that make use of its Content ID system over smaller creators. And because it’s a private system that goes beyond the DMCA, the Content ID system is under no legal obligation to comply with the DMCA’s safeguards and timelines.
ContentID certainly has some nice features — including an innovative new revenue stream for content creators. But there are significant problems with it, concerning how it handles fair use and public domain material, which serve to highlight why the idea of a “silver bullet” solution for online infringement is so problematic.