Data and the DMCA

biz+-+copyrightA group of 18 music industry organizations, representing songwriters, publishers and musicians, filed comments last week with the U.S. Copyright Office in response to the office’s congressionally ordered inquiry into the operation and effectiveness of Section 512 of the Digital Millennium Copyright Act, otherwise known as the “safe harbor” provisions of the law.

And to make a long filing short: they don’t think it’s working very well at all.

As discussed in a previous post here, people in the copyright industries have been pining for years for a do-over of the safe harbors and notice-and-takedown procedures, which they claim have amounted to a get-out-of-jail-free card for online infringement, and the music groups came prepared.

Their 70-page filing (including appendices) is chock full of colorful charts and graphs, footnoted legal arguments, and research findings, all purporting to demonstrate conclusively that the Section 512 bargain, which was intended to provide a limit on the liability of online technology providers for infringement by their uses while creating legal tools for copyright owners to get infringing content removed from the web, has been rendered null and void by advances in web technology and overly narrow interpretations of the law by the courts.

“Several factors have contributed to the failure of the DMCA to fulfill its purpose,” the groups said in their filing. “To start, Congress enacted the DMCA in 1998 when dial-up Internet speeds and static web sites predominated. Soon thereafter, individuals could be worldwide publishers of content on peer-to-peer networks and service providers began to distribute massive amounts of content uploaded to their servers. And then came along more sophisticated search engines, social networks, and an explosion of smartphones and other mobile Internet access devices. The rules for service providers and tools for content creators set forth in the DMCA proved unsuitable for this new world.

“For example, the notice-and-takedown system has proved an ineffective tool for the volume of unauthorized digital music available, something akin to bailing out an ocean with a teaspoon….Given all of these fundamental changes, a law that might have made sense in 1998 is now not only obsolete but actually harmful.”

A particular irritant, according to copyright owners, is that allegedly infringing files must be identified by their particular URL before service providers are required to take them down, according to a series of court rulings.

Once removed, however, the files are typically quickly reposted under a different URL, requiring copyright owners to send multiple takedown notices for the same work.

“Copyright owners should not be required to engage in the endless game of sending repeat takedown notices to protect their works, simply because another or the same infringement of the initially noticed work appears at a marginally different URL than the first time,” the music groups argue. “The current standard of “URL by URL” takedown does not make sense in a world where there is an infinite supply of URLs.”

The filing suggests a variety of technologies that, if implemented widely, could at least increase the effectiveness of the current notice-and-takedown regime, including:

  • Audio fingerprinting technologies to prohibit the redistribution or posting of unauthorized recordings that match previously flagged recordings or musical works;
  • Hash-matching technologies to prohibit the redistribution or posting of the exact instances of infringement previously noticed as infringing;
  • Meta-data correlations by a service provider to identify potential infringements of previously noticed copyrighted works.
  • Automatic removal and/or disabling of any links that point to a previously noticed infringement.

Alas, as the music group’s themselves acknowledge, none of those approaches, currently, is foolproof:

The type of technology used and its implementation significantly affect the scope of the content identified, the action taken after such content is identified, and its overall effectiveness. For example, for a service that implemented hash-based content identification, the number of notices of repeat infringement that were sent by RIAA declined only slightly, from 99% repeat infringements to 97% the following year. On the other hand, when a service used a more robust Audible Magic implementation designed to address master sound recordings, RIAA found that, after such implementation, no meaningful infringements of full commercially released sound recordings were hosted on that service. This meant RIAA could focus its efforts on this service to detecting and dealing with on pre-release leaks, for which content matching was not yet available.

The implementation of such technology is certainly appreciated by the Music Community but should not been seen as a panacea, particularly when, despite the availability and increased use of content matching technologies, to date services have been unwilling to address these issues on an industry-wide basis.

It seems unlikely (though not impossible) that congress would revise the law so drastically as to require online services to implement content-based filtering. And absent that, no court is going to order it at this point given the established case law.

Even if congress were to go that far, however, the industry would face a serious data problem. There is currently no universal database of recordings and associated metadata that a filtering system could reference to determine definitively whether an uploaded track is the same one as was taken down before.

The International Sound Recording Code (ISRC) lookup tool recently unveiled by SoundExchange was an important step for the industry, but ISRCs and other types of metadata can be stripped from MP3 and WAV files to make such matching difficult.

blockchain-wallet-graphic-300x200A better approach might be to leverage blockchain and other distributed technologies to permanently link digital files with their metadata.

One such effort is Mediachain, an open-source, blockchain-based decentralized metadata protocol being developed by NY-based Mine Labs.

As described by its founders, Mediachain creates cryptographically signed metadata statements about digital works which are then stored on the InterPlanentary File System (IPFS), where they become permanent and immutable. Those statements are then time-stamped on the Bitcoin blockchain, where they can be looked using an instance of the media itself. Thus, the metadata can never be disassociated from the original file, and any instance of the file can be used to recover the metadata.

I wouldn’t expect congress to get into the weeds of the InterPlanetary File System in any revision of the DMCA, of course (although there are several members who probably qualify as inter-planetary themselves). But the debate over Section 512 of the law does serve to highlight the yawning data gap that many rights-based industry face today in securing and efficiently exploiting those rights on digital platforms.

Finding innovative ways of closing that gap — by leveraging technologies such as blockchain and IFPS — will be a main of discussion at the inaugural RightsTech Summit in July in New York sponsored by Concurrent Media Strategies and Digital Media Wire. Click here to request more information about the summit or to suggest topics and speakers.

And don’t forget to visit to stay up to date on all the rights-tech news.

This article first appeared in Concurrent Media.

Post a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Get the latest RightsTech news and analysis delivered directly in your inbox every week
We respect your privacy.