Professor Peter S. Menell, Director, Berkeley Center for Law & Technology,
University of California at Berkeley School of Law
August 31, 2010

With nearly 12 years of experience, we are now in a better position to assess the effects of the Digital Millennium Copyright Act (DMCA) Section 512 safe harbors.  At the time that Congress passed the DMCA, there was substantial concern that the risk of crushing liability for indirect copyright liability would discourage the growth of the Internet.  At the same time, copyright owners feared that blanket immunity for online activities would lead to rampant unauthorized distribution, destroying their ability to exploit their works in both offline and online markets.

Section 512 attempts to balance these competing interests by affording online service providers (OSPs) immunity from monetary liability for transmission, caching, storage, and linking so long as they satisfy various requirements.  Under Section 512(c), OSPs storing content at the behest of users must expeditiously remove such content upon notice by the owner that the material being hosted is infringing – hence the term “notice and takedown.”  Furthermore, under the “red flag” provision, the OSP loses the protection of the safe harbor if it is “aware of facts or circumstances from which infringing activity is apparent” and does not take action – even without formal notice by the content owner.  Thus, the DMCA allocates primary policing responsibility to content owners.  But OSPs bear responsibility to remove content even in the absence of formal notice to the extent that they are aware of “red flags.”  During the past two years, district courts have begun to address these issues in the context of user generated content (UGC) websites.

So how is this regime playing out?  What are the messages being received in the online marketplace?

Last Monday, I had the opportunity to consider these questions on a panel entitled “The Internet and the Media – After the First Wave, What’s Next?” at the Technology Policy Institute’s Aspen Forum.  A webcast can be found here.  The panel included a senior product counsel for YouTube, the CEO of a music Internet start-up called Grooveshark, the head of the Recording Industry Association of America (RIAA), and two economists who have studied the impact of the Internet on content industries.  To borrow from Clint Eastwood’s oeuvre, the panel illustrated what I would call the “The Good, the Bad, and the Ugly.”

In my view, YouTube as it now operates is a shining example of the “good.”  Backed by venture capitalists, YouTube launched in November 2005 as a video-hosting and streaming website.  Anyone could upload videos and reach a vast audience.  This provided a great vehicle for families to post their home videos, politicians to post their messages, and budding filmmakers to post their works.

But YouTube also attracted unauthorized motion pictures and television shows, such as Comedy Central’s “The Daily Show with Jon Stewart.”  When Viacom notified YouTube of specific infringing videos, YouTube promptly removed them.  But as YouTube’s popularity grew and the supply of videos expanded, the notice-and-takedown regime exacted a substantial and growing policing price – technically referred to as the “whack-a-mole” problem.  Shortly after Google acquired YouTube for $1.65 billion, Viacom sued for copyright infringement, alleging among other things that YouTube ignored red flags and hence lost its Section 512 liability shield.

Over the next year, YouTube cleverly solved the whack-a-mole problem.  It engineered a “ContentID” system that proactively screens uploaded videos to determine whether they match a vast database of protected works and, if so, whether the owner of such works has pre-approved its uploading – typically with sharing of advertising revenue with YouTube.  This technology has substantially reduced the level of infringing content and the costs of policing.  It has also created the foundation and technical infrastructure for collaboration between YouTube and many content enterprises.  As a result, Viacom dropped its prospective relief requests and the litigation focuses only upon YouTube activities prior to the implementation of ContentID.  Whether YouTube was motivated by the risk of liability or not, the outcome appears to be a real win-win for YouTube and content creators.

But did the red flag provision of the DMCA require YouTube to develop proactive filtering?  That is a question that the Second Circuit will address in the coming year.  Perhaps YouTube would have developed and implemented ContentID without the risk posed by Viacom’s lawsuit.  At this stage of the case, the district judge has answered this question in the negative.  Which brings me to Grooveshark.

Although less well known than YouTube, Grooveshark is quickly developing a large base of Internet users.  And it is not difficult to see why.  Grooveshark offers customers the ability to stream a vast collection of the most popular recorded music on demand for free.  And for the cost of $3 per month or $30 per year, one can stream all of the music to your mobile phone and run Grooveshark’s VIP desktop platform.  The celestial jukebox has arrived.  But there’s a problem.  Grooveshark lacks authorization to store or distribute most of the music stored on its servers and flowing to its customers.

So where does Grooveshark get the music?  Like peer-to-peer systems, Grooveshark scans a share folder on its users’ computers.  Unlike peer-to-peer systems (and far more questionable), Grooveshark uploads these files to its servers.  It considers these files to be user generated content.  And taking an aggressive interpretation of the Section 512 safe harbor, it believes that it is not required to remove these files unless and until it receives a specific infringement notice.  In essence, Grooveshark has created a system that imposes tremendous whack-a-mole costs on copyright owners and hopes to use this cost as leverage to extract favorable licensing arrangements.

And the strategy has had modest success.  Grooveshark has entered into some licensing arrangements with independent labels looking for exposure.  It has also entered into a licensing deal with EMI, one of the four major record labels.  Anyone following music industry news, however, understands that EMI is heavily in debt – which makes it more amenable to leverage.  Grooveshark has snared the weak antelope in the herd.

In the interests of preparing for last Monday’s panel, I visited Grooveshark’s website to see what was available.  I quickly found all of my favorite songs and recording artists.  The interface allowed for on demand listening – something that the compulsory licensing provisions of the Copyright Act do not provide for.  But then again, Grooveshark isn’t operating under a compulsory license.

Knowing of the EMI deal, I searched for Beatles sound recordings on Grooveshark.  EMI manages the Beatles catalog and it is well known that the Beatles do not authorize Internet access to their music.  You won’t find the Beatles hits on iTunes or Rhapsody.  And you won’t find them on Grooveshark either.  How can that be?  Apparently, no users of Grooveshark have Beatles recordings in the share folder or, more plausibly, Grooveshark proactively screens them – pursuant to its agreement with EMI.

It would seem sensible that Grooveshark would provide such screening for all or even just many unlicensed commercial artists.  In fact, Audible Magic and other vendors provide the tools for such pre-screening.  But that would cost Grooveshark money.  And more importantly, it would cost Grooveshark most if not all of its user base.  And why bother if the DMCA gives you cover?

The DMCA puts content owners in a tough situation.  They can: (1) spend substantial resources whacking moles with little hope of seriously reducing the problem; (2) sue Grooveshark for copyright infringement and risk a broad reading of the DMCA safe harbor; (3) sue Groovershark under state law for infringement of pre-1972 sound recordings, as Universal Music Group has done, and open a Pandora’s box of unprecedented issues; or (4) do nothing – and watch further erosion of the music marketplace.  None of the options are very attractive.

In my view, Grooveshark reflects the cynical and “bad” approach to the DMCA safe harbor.  Grooveshark does not offer breakthroughs in technology.  It is taking an aggressive approach to the safe harbors to cannibalize the online music marketplace.  The DMCA notice-and-takedown regime provides the leverage for Grooveshark.  It sees the lower court decision in Viacom v. YouTube as an invitation to cannibalize and leverage.

So that brings me to the “ugly”: the inability of just about everyone with a stake in this debate to discuss the growing pains, shortcomings, and anachronisms of the DMCA safe harbors forthrightly.  As reflected in the views of our industry panelists, the DMCA safe harbors are either perfect and prescient or dysfunctional.  Some of the answer lies in how the courts interpret the red flag provision.  The one safe prediction that I can offer is that the positions will reverse if the Second Circuit reverses.  It should have been clear to everyone on the panel that neither Web 2.0 nor peer-to-peer technology were even on Congress’s radar when it passed the DMCA, and that the statute is not well-crafted to confront the modern challenges.

In my view, Congress would have been prescient if it provided enough affirmative nudge for UGC companies to develop proactive filtering technologies as the Web 2.0 took flight and filtering became more feasible and cost-effective.  Innovation has an important role to play in achieving the balance that the DMCA sought to achieve.  Unfortunately, if the red flag provisions amount to nothing, there will be little incentive for UGC sites to make the effort.  That said, a risk of billion-dollar statutory damage remedies – as alleged in the YouTube case – makes little sense in the absence of billion-dollar harms.  Congress lacked prescience in ramping up the statutory damage range in 1999 without recalibrating for a Web 2.0 world.  Thus, there are good reasons to update the DMCA safe harbor and statutory damage provisions.