|  | ||||||||||||||||||||
| | ||||||||||||||||||||
| Chilling Effects Clearinghouse > Copyright > Weather Reports > YouTube Upgrades its Takedown Challenge Procedure for ContentID-Triggered Video Removal | 
|
|  YouTube Upgrades its Takedown Challenge Procedure for ContentID-Triggered Video RemovalAdam Holland, Chilling Effects, October 04, 2012 Abstract: Until quite recently, ContentID, YouTube’s internal monitoring software to detect material that (theoretically) infringed copyrights, had little or no appeals process. This was frustrating for many YouTube users, leading some to stop using YouTube altogether and others to write long and thoughtful "Dear YouTube" letters, in hopes of creating positive change. Until quite recently, ContentID, YouTube’s internal monitoring software to detect material that (theoretically) infringed copyrights, had little or no working appeals process. This was frustrating for many YouTube users, leading some to stop using YouTube altogether and others to write long and thoughtful “Dear YouTube” letters, in hopes of creating positive change. See some of Chilling Effects' prior posts on this subject here. But yesterday, YouTube users everywhere were given at least some reason to celebrate. Whether it was the gradual accumulation of user complaints, the thoughtful letter above (posted only two days before YouTube’s announcement!) or the egg on YouTube’s face from the recent erroneous takedowns of the Democratic National Convention’s livestream as well as public domain NASA footage, YouTube announced fairly substantial revisions to their Content ID policy, specifically with respect to posters challenging takedown of material. What’s critical to remember in this context is that the ContentID takedowns are completely internal to YouTube, a corporate policy decision that YouTube made. The takedowns, in the form they have under ContentID, are not required by law. Of course, this means that their details are not spelled out by law either. The closest relevant law is the Digital Millennium Copyright Act, the DMCA. Its notice and takedown provisions are well known to regular visitors to Chilling Effects: Under the DMCA, a copyright owner who believes that their work is being infringed notifies the entity hosting the material. Here, YouTube. In order to take advantage of the DMCA’s “Safe Harbor”, YouTube must take the material down. Critically, no evaluation of the legal validity of the claim by YouTube takes place. Next, if the poster of the allegedly infringing material disagrees, and thinks that the material in question does not infringe, they may submit a counternotice to YouTube. Upon receipt of a counternotice, YouTube replaces the material in question, after waiting a statutorily required 10 days. At that point, the putative copyright holder/ original complainant’s only recourse is legal action – most likely filing for an injunction. This procedure has received a great deal of criticism due to its imbalance--the ease with which material can be taken down compared to the difficulty in getting it replaced, no perjury penalties for the complainant’s original claim--but there is a way, at least, for a poster whose material has been mistakenly taken down to get that material replaced. And it’s at least arguable that at the third stage, the imbalance then swings the other way, given that if a poster goes to the trouble of filing a counternotice under penalty of perjury, 17 U.S.C 512(g)(3)(C), the only recourse for a rightsholder is the dramatic and resource-intensive step of pursuing legal action. Contrast this with YouTube’s Content ID policy up until now. First, the ContentID software took material down without ever having received a complaint from the copyright owner, instead relying on automatic software matching with content uploaded by "content partners", content which is apparently not vetted to see if the uploader actually possesses copyright in it. (See here for an interesting examination of what the matching software might be looking for.) That is, no proof of copyright ownership. much less registration of copyright, is required. So the deck is somewhat stacked right away. Second, although a poster could dispute the takedown, which would result in the video going back up the putative copyright holder, who had uploaded the reference material for which a match had been found, could simply “reject” the dispute, making filing a dispute an utterly toothless power for posters. Interestingly, YouTube’s explanation of the ContentID process doesn’t mention the possibility of this outright rejection. A nice flowchart can be found here. But most saliently, in contrast to other Google-related takedowns YouTube provided little or no information about the ContentID takedown process, not even which takedowns were due to DMCA and which to ContentID. Further, as YouTube admits, YouTube users whose videos were taken down , even by clear mistake, were "left with no recourse for certain types of Content ID claims (e.g., monetize claims).” and could address the Kafka-esque situation only through some sort of “know the right person” personal appeal, which of course have no certainty of success. YouTube was previously aware of the problem but until now hadn’t offered a solution. [See also this prior post by the same author, in a different role.] The new procedures map much more closely to those of the DMCA, and in fact, reference it directly, and make it an inextricable part of the procedure, On top of that, YouTube’s blog post seems to hint that the ContentID software will not be quite as black and white as it has been in the past, and may now have third category,a gray area of sorts, whereby a video flagged as infringing by ContentID will then be sent to a human being for review. This is heartening, because it provides at least an opportunity to bring fair use into the evaluation process again, which is something software is unlikely to ever be able to do, no matter how sophisticated. Judges find it challenging enough! However positive a step this is, does this improvement by YouTube go far enough? Content ID still polices YouTube, taking material down if it matches uploaded reference material. Users who get three strikes are still kicked off YouTube, although a user can erase a strike by going to “Copyright School” itself a topic worthy of its own Report. What’s more, there are limits to the number of times a poster can appeal a rejected dispute. It’s unclear if a strike will be erased if the claimant later releases the claim. Although more could have been done, such as creating a whitelist to balance the library of reference materials uploaded by content partners; allowing individual users, not just large corporate content holders, to submit reference materials; or more explicitly recognizing the role and scope of fair use--a knotty problem for almost any automated takedown scheme--on balance, this is a laudable decision for YouTube. Not only have YouTube made their internal procedure more even-handed by giving posters clear avenues by which to seek redress, they have also linked their internal procedures to the existing, and, for whatever its faults, well-known, external procedure of the DMCA, thereby moving a debate over the nature of particular content out of YouTube’s private sphere and into the unquestionably more transparent milieu of DMCA notice and takedown. Participants on both sides now will have a much clearer idea of what is going on with respect to their particular video, and the public in general will ideally get a more robust and nuanced look at the scope of YouTube’s internal takedowns, analogous to Google’s own transparency report.
|
|
|
|
Recent C&Ds