I’ve previously addressed, a few times now, the problem with relying too much on automated systems to identify infringing files that are then targeted with DMCA take-down notices. The latest widely-reported incident of this kind is even more unfortunate in some ways, because the victim of the unwarranted take-down notice also happens to be a fairly vocal critic of the DMCA: author, blogger and activist Cory Doctorow.
The issue is pretty simple; in an effort to remove files (and links to those files) that infringe on its popular TV Show Homeland, 20th Century Fox also mistakenly issued take-down notices relating to Doctorow’s novel of the same name.
As you can imagine, Doctorow is less than thrilled by this development. While he accepts that it was done by mistake, he’s still flummoxed by that error, and understandably so.
“It’s clear that Fox is mistaking these files for episodes of the TV show ‘Homeland,’ Doctorow wrote. “What’s not clear is why or how anyone sending a censorship request could be so sloppy, careless and indifferent to the rights of others that they could get it so utterly wrong.”
I’m a bit more sympathetic than is Doctorow where mistakes like this are concerned, mostly because I reserve most of my sympathy where piracy is concerned for the rights-holders whose work is being shared willy-nilly by large numbers of people who show the copyrights owned by those rights-holders no regard whatsoever.
When you deal with thousands of infringements at a time, which I’m sure Fox is with respect to Homeland, it quickly becomes unfeasible to review each possible infringement individually, so you create programmatic short cuts designed to reduce the “leg work,” so to speak.
In the case of DMCA Force, we use our content spiders look at file names, link text and descriptive text that might accompany the files, among other things, to help us sort potential infringements based on the likelihood of actual infringement. Once the results have been sorted, we look for clues to use in the future that could indicate infringement (or just as useful, non-infringement) and adjust our system, accordingly.
Although we’re quite confident in our software, the last line of defense against this sort of error, for us, is our individual analysts. We rely on the hardest part of a human being to simulate in the context of artificial intelligence — subjective, individual judgement — in order to reduce the likelihood of errors to the lowest possible level.
Even with all that effort we put in to prevent mistakes, I’m sure we’ll make some along the way — which is why I’m inclined to go softer on Fox in this situation than is Doctorow.
The reason today’s post title ends with a question mark is that I’m not entirely sure this particular take-down notice error is the result of over-reliance on automation in the service provider’s system. If it isn’t, then the error becomes a lot harder to explain. If the URLs in the notice at question were assembled by hand, it gets more difficult for me to fathom missing Doctorow’s name, right there in the URL itself.
Having said all the above, the real problem here isn’t Fox; the problem is all those people who feel they have the right — and possibly even an obligation — to “share” that which is not theirs. No amount of bogus take-down notices justifies the total disregard for intellectual property that many Internet users display these days, and it’s important that we don’t forget that, even as we chide rights-holders for their own foibles.