Future of Music Coalition (FMC) respectfully submits the following reply comments to the Copyright Office inquiry on the safe harbors and notice-and-takedown provisions within 17 U.S. Code § 512. 

Future of Music Coalition (FMC) is a Washington D.C.-based nonprofit organization supporting a musical ecosystem where artists flourish and are compensated fairly and transparently for their work. FMC works with musicians, composers and industry stakeholders to identify solutions to shared challenges. We promote strategies, policies, technologies and educational initiatives that always put artists first while recognizing the role music fans play in shaping the future. FMC works to ensure that diversity, equality and creativity drives artist engagement with the global music community, and that these values are reflected in laws, licenses, and policies that govern any industry that uses music as raw material for its business.

Our comprehensive research into artist revenue streams[1] along with longstanding support of access to audiences in broadcast[2]and digital media[3]has given us a pragmatic, creator-centric view of copyright enforcement, creative expression and technological evolution. FMC’s governing board is a microcosm of the constituencies that we serve, including performing artists, songwriters, music managers, independent labels, data analysts, music licensing experts and creative technology developers. Our daily interactions with working musicians and composers across genres and disciplines, along with our history of translating marketplace and policy developments for our creator allies are germane to the issues put forth by the Office in this request for comments. We appreciate the opportunity to participate in this study and hope that our perspectives, which are informed by consultation with music creators and their teams, will prove useful to your analysis.

Though our expertise is primarily focused in one sector, our comments reflect a broad and participatory understanding of content and technology markets. We have zeroed in on the questions we feel best equipped to answer based on our mission statement and experience; our responses are ordered sequentially, but non-consecutively. 

1. Are the section 512 safe harbors working as Congress intended?
It is rarely a simple matter to ascertain the will of Congress, even when legislation under examination is as recent as two decades ago. However, basic assumptions can be formed through a non-ideological plain reading of statute, along with an examination of the legislative record, both of which lend themselves to the conclusion that Congress intended to take a light touch with information exchanged on the Internet, based in their presuppositions that such an approach would allow for greater innovation across the burgeoning worldwide web. If the above is indeed reflective of Congressional intent, then in many ways, the safe harbors have worked well. However, such a conclusion ignores many aspects of the contemporary digital marketplace that Congress was unlikely to have contemplated, much less fashioned an anticipatory response to.

First, the incredible growth of the web also unleashed a tsunami of infringement, with the notice-and-takedown provisions of § 512 serving as both port in the storm and active fault line. There is little doubt that safe harbors have contributed to the massive growth and adoption of the Internet, making possible many legal, licensed services that many music creators use every day, yet this growth has also brought scale to unauthorized access, reproduction and distribution of copyrighted works. Google’s Transparency Report for the month of February 2016 pointed to some 80,510,012 takedown notices processed, averaging out at some 100,000 notices processed per day.[4] Of course, this represents a very small fraction of what the search engine indexes, and this particular report does not include the company’s other products. Still, it is easy to understand content creators’ reaction to these numbers, especially smaller rightsholders who lack the financial and human resources to respond to infringement at this scale and volume. Some feel that responding to individual instances of infringement is generally a lost cause, particularly on those sites or services where even finding the required information about a site’s DMCA compliance is a time-consuming chore that requires wading through pages of banner ads and pop-up windows. Individual creators and independent music companies alike have also expressed deep frustration at the fact that compliance with legitimate takedowns does not address the problem of unauthorized reposts of the same material. Furthermore, court interpretations of statute do not support the obligation of a service to monitor its site or user activity for infringements, as the Office itself notes with regard to court opinion on § 512(m).[5] However, there is a standard technical measure exception described in subsection (i). This is important, as this subsection describes a consensus approach to the“technical measures that are used by copyright owners to identify or protect copyrighted works.”[6]These measures, coupled with good-faith, multilateral oversight, may prove the most effective approach to living up to the spirit of the statute and the incentives in the Progress Clause of the United States Constitution.

Smaller digital service providers (DSPs)—including nonprofits and emerging user-focused applications—may have difficulty processing even a relatively small number of requests due to incomplete, incorrect or invalid information provided by rightsholders via automated or human-generated notices. The standardization of forms, templates, and requirements may help, though smaller service providers that are not in a position to afford or implement fingerprinting technologies to identify potentially infringing material may find obligations beyond that which statute provides to be burdensome. However, we reject the notion that any voluntary standardization measures or efforts to implement sector-wide best practices are inherently burdensome to DSPs. We recognize that a diminishing of limitations on liability may indeed impact investment in potentially useful digital services, which means that enforcement protocols must work for small-to-medium sized enterprise (SME) in both the technology and creative sectors. We need to make it easier for all parties to do the right thing.

Our reading of § 512 suggests that Congress sought not to incentivize infringement, but rather to establish a system to address it that balances the interests of stakeholders. In subsequent sections, we will examine why we believe that the law is clear, but aspects are underrepresented in practice. We appreciate the Copyright Office inquiry in part because it provides an opportunity to discuss statutory obligations in a way that might inspire parties to develop mutually agreeable best practices and high standards of compliance.

It must be noted that a healthy digital ecosystem requires more than just properly calibrated enforcement. Looking only at the music sector, it is obvious that there are tremendous issues in data standards, data accessibility and data transparency (both public and between contractual parties) that must also be resolved. Going further, our industry remains handicapped by legal disparities, loopholes and anomalies within the music licensing and royalties regime. It would be prudent to not lose focus on these matters while considering how best to address the burden of enforcement online, as inadequate policy on one side leads to the perpetuation of problems on the other. It is our view that having multilateral discussions out and in the open is preferable to legislative skullduggery. We believe that best practices and voluntary agreements among stakeholders can help honor the law’s broader purpose of protecting creators and rightsholders while shielding constructive technologies from inappropriate precedential weighting or overly broad statutory mandates. This is approach isn’t just pragmatic or good for business; it is in keeping with the constitutional imperative to “promote the Progress of Science and the useful Arts…”[7] What must now be encouraged is an alignment of incentives and good-faith problem-solving. It is our hope that this study helps to incentivize stakeholders in this regard.

 3. How have section 512’s limitations on liability for online service providers impacted the growth and development of online services?

Limitation on liability has clearly benefitted the development of services that otherwise may have had difficulty securing investment capital against risk of damages, or maintaining specialized personnel necessary to ensure that no user-upload content contains potentially infringing material. It is fair to say that many of the services that enjoy limits on liability under §512 are useful to those in the creative sector. Musicians were among the earliest adopters of Twitter; in fact, the top three most-followed accounts on the service are musicians Katy Perry, Justin Bieber and Taylor Swift, all of whom are ahead of President Barack Obama.[8] Artists and labels that cater to smaller audiences also use Twitter and other services to promote their work, advertise concerts and special events and cultivate brand affinity. Facebook, Instagram (acquired by Facebook) and Tumblr (owned by Yahoo!) likewise provide environments where artists can engage with fans and maintain interest in a crowded music marketplace. Services like Twitter-owned Vine—which publishes user-generated 6-second videos—has led to major label signings of such artists as rapper Bobby Shmurda, folk-pop duo Us and 16 year-old singer Shawn Mendes. It is possible that none of these artists would have been discovered without access to §512-eligible social platforms.

Safe harbors have also made things easier for dedicated music services where artists make their recordings available for purchase and streaming. The popular direct-to-fan platform Bandcamp is well-respected in the music community due to the fact that artists can build out their own pages and sell music at flexible price points. Bandcamp likely does not receive a high-volume of takedown requests, but the service may have never gotten off the ground without limited liability protections. Unfortunately, it is abundantly clear that not all services, whether they are social platforms or access points to specific media, operate in good-faith with regard to §512. The notorious and now-defunct on-demand streaming site Grooveshark, for example, made repeated claims that it enjoyed safe harbors due to the fact that users posted the infringing material (full length, copyrighted audio tracks implicating both sound recording and composition exclusivities); ultimately, courts decided that the activities of executives and staff rendered the service ineligible for liability shields under statute. The problem here is that Grooveshark had already exploited the idea of safe harbors to build and monetize a high-traffic site that served no other purpose than to stream 1 billion songs per month to some 20 million listeners, all while misleading consumers about its legality. Some of this music was at least partially licensed; the vast majority was not. Legal recourse is not always available to all creators and rightsholders whose works have been infringed, whether or not a court finds a service eligible for safe harbors. In fact, the high expense of litigation means that only the best capitalized music companies can afford to bring action against DSPs that do not meet statutory requirements. To us, this is one of the most overlooked aspects the safe harbor debate. It is our firm belief that remedies should be available to all categories of author under §512; if lawsuits and high-volume takedown requests are not options for small-to-medium sized rightsholders, there needs to be an examination of how to make the system work better. Clearly, there exists an inflection point where the burden and cost of direct licensing and its inherent chilling effect on innovation must be balanced against services that reach that inflection point and can compete on a rate and content playing field against services that do obtain licenses based on a willing buyer/willing seller rate negotiation. Solving these issues may require a novel approach to content licensing for certain categories of service, which we will explore in a subsequent section of our comments. On the other side of the coin, if there are, as some research suggests, a large number of incomplete or incorrect notices sent to some smaller or emerging services and/or the threat of ineligibility due to uncertain legal status, it may frustrate the development of technologies that ultimately prove important to content creators. The challenge is to achieve a better balance without creating additional burdens on small-to-medium-sized enterprise in both technology and music.

4. How have section 512’s limitations on liability for online service providers impacted the protection and value of copyrighted works, including licensing markets for such works?

To the extent that section 512 as currently implemented results in consumer access to unlicensed copies of copyrighted works, it certainly can have a negative impact on both the protection and value of this material. When creators and rightsholders are competing with free, there can be an inescapable and persistent downward pressure on pricing. This puts rightsholders at a disadvantage in licensing negotiations, and can mean that promising services that attempt to license are unable to do so due to the imposition of high capital risk demands by rightsholders who are reluctant to play “wait and see” while unlawful consumption escalates. At the same time, to the extent that section 512 as currently implemented results in new avenues for selling and licensing music and establishes profitable markets, it can create positive impacts as well. Common-sense improvements and best practices in how statutory requirements are implemented could result in more positive impacts and fewer negative impacts.

It’s worth noting that artists or rightsholders may come to sharply different conclusions about unauthorized uses of their works, dependent on such factors as genre, scale, audience, career stage, demographic, or other factors. This is unsurprising, as music is a highly specialized field; our 16 years’ worth of qualitative and quantitative data gathering has time and again demonstrated that musicians and composers are not a monolithic group. As a general rule, when two creators report divergent outcomes from the same marketplace, technological or policy development, it’s best to assume both are accurately reporting their experience and avoid lionizing any approach as the “correct” way to do business in the digital age. A diversity of marketplace approaches fuels a diversity of creative expression. Because copyright law is intended to serve all categories of authorship, as well as the public interest, successful policy interventions and best practices should help address the grievances of those artists and rightsholders who report negative impacts of a generalize loss of control over the use of their work, while allowing those who report positive career impacts from third-party uses to continue to encourage such activity. Similarly, although we do not have data on the developer community, our interactions with founders and principals of such companies as PledgeMusic, CD Baby and Bandcamp (to name a few) indicate that it is unwise to paint the tech sector with a single brush. Many are trying to solve problems in the digital content sector and deliver value to creators, rightsholders and users. Where there are difficulties aligning incentives and balancing obligations, we believe that agencies such as the Copyright Office and IPEC may have a positive role to play.

The “infringe now, license later” approach to digital distribution has been a particularly vexing issue for rightsholders. On one hand, some artists and rightsholders desire to make their work visible on platforms where there are clearly a large number of highly engaged fans. On the other, emerging platforms can attract massive private investment sums, allowing them to reach scale before obtaining licenses from rightsholders for copyrighted works that appear on the service as third-party uploads. By the time negotiating for a license finally happens, the service has scale and platform popularity on its side as negotiating leverage; rightsholders, particularly SMEs, are disadvantaged and may seek additional comfort in the form of larger cash advances, minimum monthly guarantees, Most Favored Nation status, breakage and partial ownership in the form of equity. Our point is not necessarily to criticize any one aspect or the entirety of digital deal structures, but rather to illustrate how insecurity around safe harbors and liability can produce outcomes that may end up affecting the viability of services that tend to over-index independent, niche and emerging music.

Different authors or rightsholders may attempt to navigate these dynamics in different ways. Soundcloud, which built its reputation on an elegant audio player and core userbase comprised of DJs, remix artists and electronic dance music (EDM) producers, is the latest service to go from unlicensed to legitimate, evolving fundamental aspects of the service along the way.[9] While moves towards mutually agreeable licensing terms should be applauded, aspects of how these changes have been implemented have not sat well with many users, including some music creators. Before Sony Music licensed to Soundcloud, it required the service to remove the tracks it owns; the service complied with this request using automated identification technology. Electronic artist Kaskade, who released music on a Sony subsidiary and credits his success in a large part to the DJ and producer community on Soundcloud, took issue with the fact that his work was being pulled down from the environment that aided in his career development and likely got him noticed by a label in the first place.[10] It is important to note that Sony Music makes decisions about how it licenses its catalog and repertoire based on a number of factors, and that an individual artist under contract with the company is unlikely to have the leverage to affect these decisions. Still, the incident highlights the tensions between protecting rightsholders and serving a universe of artists who have diverse and highly specialized roles in the space. Meanwhile, other music creators now express frustration that Soundcloud is launching a new subscription service without clearly indicating how payment is supposed to work.[11]

We recognize that not all artists and composers on a potentially eligible §512 service are affiliated with rightsholders who have made direct deals outside of the safe harbors. In such instances, depending on the nature of the service, aggregators like CD Baby, Tunecore, and Distrokid may serve as middlemen for many of the wholly independent music creators who desire to participate. In the case of Soundcloud, this had no bearing until very recently. From the launch of the service in 2007 to now, the issue of safe harbor eligibility for that particular remains unresolved as a purely legal matter. That having been said, it is possible that the mere existence of the notice-and-takedown regime encouraged the service to pursue negotiations with rightsholders, as it did not want to have to devote so much focus to responding to notices as it established a business model for monetization. For copyright owners, however, safe harbors may create conflicting incentives. Music companies are in part motivated by whether a platform has a critical mass of users, but would no doubt would prefer to do direct deals under a mutually acceptable revenue model rather than resign themselves to the Sisyphean task of sending multiple, repeating notices.  Unfortunately, in the world of technology, revenue models aren’t always clear to the coders of a service or application at early or even mid stage, as user adoption and engagement often informs what commercial path a platform should focus on developing. Though this fuzziness may be a boon to certain kinds of innovation, it puts artists and rightsholders at a distinct disadvantage, particularly on user-upload and/or social platforms. We believe that ameliorating these tensions is possible if stakeholders consider the broader intent of the statute as well as some of its overlooked provisions.

With regard to the exploitation and monetization of copyrighted works, it is obvious that a great many creators and rightsholders believe that safe harbors specifically have contributed to an overall devaluation of music. While we are not aware of longitudinal research supporting or contradicting this view, we are certainly able to entertain the idea that some technology companies are in a position to use the safe harbors as a workaround to grow their businesses without an immediate plan to compensate the owners of copyrights that attract users to a service or application. However, we would strongly encourage the Office to consider the multiplicity of contributing factors to devaluation—from pre-Internet commodification due to rampant media consolidation to post-iTunes album unbundling to the incredible competition for mindshare that any expressive work faces in an era of content abundance and near-zero marginal cost reproductions. With regard to §512, we would push back against any service or application that uses safe harbors to compel below-market compensation, or that sets a ceiling for negotiated or statutory rates.

Additionally, to the extent that safe harbors currently result in the widespread availability of unlicensed copies of creative works, this dynamic can create new imperatives to license entire catalogs under one-size-fits-all terms. While creators’ and rightsholders’ opinions on the rise of “full-catalog-on-demand” access-based business models vary widely, such models are frequently criticized by artists for how they tend to reward works that achieve global scale and repeat listenership. Such works are not necessarily created with a goal of mass appeal and may not entice frequent repeat listening, which in turn can heighten a winner-take-all dynamic. There may even be an unspoken premise that a licensed use at a low rate is superior to an unlicensed use with no compensation. If creators and rightsholders—particularly SMEs—enjoyed better access to the remedies spelled out in §512, the music community may not feel obligated to participate in “one size fits all” business models, thereby encouraging a greater diversity of digital platforms more specifically calibrated for niche audiences and diverse traditions.

The purpose of safe harbors is to encourage innovation while providing a course of action for rightsholders whose works have been infringed. Yet we frequently hear complaints from smaller labels, publishers and music creators about the low payments on YouTube, sometimes coupled with frustrations about access to its proprietary detection technology, Content ID, including monetization options. The pervasive use of non-disclosure agreements adds a level of obscurity; it can be difficult for small independent labels and artists to get clear information from digital distributors or third-party partners who provide access to content management systems (CMS) regarding the specific degrees of control rightsholders and creators actually have over the use of their work.[12] Whether such NDAs are implemented at the request of rightsholders or digital services, their ubiquity encourages mistrust and confusion. We would encourage all music services and rightsholders to be more transparent and flexible with regard to terms and placement options, and take steps to ensure that access to detection technology is not dependent on accepting uniform licensing terms. Many copyright owners and music creators will continue to make their works available on both free and subscription tiers if they see value. Some music managers report that access to their artists on a platform that boasts 1 billion active users can deliver positive externalities in other areas, such as live performance, brand deals and synchronizations. Positive externalities, however, must not be used as a crutch; the maxim, for instance, that the promotional value of radio outweighs the need for a performance royalty to sound recording copyright owners and performers is an example of this thought process going awry. In a §512 environment, rightsholders have grown wary of the idea that massive businesses can be built on the backs of copyright creators and owners. It is a truism in the music industry that everyone wants to be paid more, from bedroom producer to senior label executive. However, the distribution of opportunity in digital is neither equivalent nor necessarily democratic. Stakeholders would do well to collaborate on licensing and enforcement frameworks that allow the largest degree of flexibility to creators and rightsholders while not diminishing the value that services like YouTube deliver to a wide category of content creators, some of whom have established successful channels by producing videos that would otherwise have had difficulty finding a market in traditional media environments.

9. Please address the role of both “human” and automated notice-and-takedown processes under section 512, including their respective feasibility, benefits, and limitations.

When the safe harbors were first established, the technology to detect copyrighted works within services and applications was in its infancy. However, §512(i) does provide for a certain approach to automated detection:
  

(i) Conditions for Eligibility.—

  (1) Accommodation of technology.—The limitations on liability established by this section shall apply to a service provider only if the service provider—

        (A) has adopted and reasonably implemented, and informs subscribers and account holders of the service provider’s system or network of, a policy that provides for the termination in appropriate circumstances of subscribers and account holders of the service provider’s system or network who are repeat infringers; and

        (B) accommodates and does not interfere with standard technical measures.

  (2) Definition.—As used in this subsection, the term “standard technical measures” means technical measures that are used by copyright owners to identify or protect copyrighted works and—

        (A) have been developed pursuant to a broad consensus of copyright owners and service providers in an open, fair, voluntary, multi-industry standards process;

        (B )are available to any person on reasonable and nondiscriminatory terms; and

        (C )do not impose substantial costs on service providers or substantial burdens on their systems or networks.

Repeat infringer policies are not explicitly spelled out in statute, which provides an opportunity for productive dialog about what a standardized policy might look like, or at least a renewed commitment on behalf of stakeholders under a voluntary agreement framework to make clear corporate policies regarding repeat infringers with appropriate compliance. Crossing this off the list of persistent annoyances is mostly a matter of taking the time to identify provisions that are mutually agreeable and enforceable. For example, it might be a good practice to give rightsholders the option to remove an unauthorized copy of a work without incurring a “strike” against the uploader—if and only if the rightsholder chooses not to impose this penalty. While it’s important to give rightsholders a means of addressing the very real problem of repeat offenders, our conversations with artists and independent labels tell us that in practice, some artists and labels may leave successfully detected infringing copies online, simply because they don’t want to punish a fan who may not have known any better. Popular-level understanding of how copyright works online is generally poor, as evidenced by the preponderance of meaningless “disclaimers” appended to uploads on UGC sites (e.g.: “I do not own this! No copyright intended!”). Giving creators a means to enforce their rights without penalizing a clueless fan might be a small step towards better outcomes.

The matter of automated detection is more difficult to reconcile. Other filers will no doubt highlight the difficulties that can be encountered when a machine fails to recognize a potentially fair use, which typically comes with a laundry list of grievances regarding false positives and expressive rights denied. Some of these concerns can be addressed through implementation of technological tools, such as emphasizing the detection of complete works, which likely have the biggest potential for negative commercial impact and are least likely to be deemed “fair.” While we believe that fair use is a vitally important principal and fundamental to any serious understanding of US copyright law, there is also need to focus on the practical matter of what tools are necessary to streamline the notice-and-takedown process, and how uncontested takedown notices that satisfy informational requirements might be aided by technologies that match the actual digital content to the claim. To that end, rightsholders must leave behind any remaining parochial and protectionist attitudes with regard to ownership data and deploy all recorded music tracks with standard numeric identifiers that accompany both the sound recording and the underlying composition. Having advocated for better music metadata for more close to two decades, FMC and our allies in the music community continue to make inroads with regard to the data quagmire. The details of these efforts are beyond the scope of our comments, though we would emphatically note that any and all detection technologies and human oversight policies would be tremendously aided by even basic commitments to maintaining accurate and authoritative information on rights ownership.

When we examine at the current state of automated detection, we again see a lopsided marketplace of have and have-nots. Beyond the much-discussed Content ID, we are familiar with detection tech vendors such as Audible Magic Vobile, and Audiolock, though there are surely other companies in this space providing similar services. However, many of these technologies are frequently inaccessible or unaffordable to both small content creators and tech developers. A small independent record label may only be able to afford a third-party service for its most high-profile anticipated releases, and then only for a short window before and after its release. An emerging technology with a strong UGC backbone that wants to live up to its DMCA obligations likely needs to keep detection tech engaged for as long as the platform remains publicly accessible. We recommend that either the Copyright Office or IPEC convene a targeted group of stakeholders to address this and other issues around automated detection. We envision this group to include music creators or their representatives, detection technology vendors, and major and independent record labels and publishers and software and application developers and consumer advocates. We believe that a tighter focus on one particular issue—in this case how best to effectuate takedowns while maintaining important exceptions—will potentially lead to better outcomes than large group convenings. The goal, beyond offering the opportunity for discussion among parties who rarely interact, would be to arrive at best practices and action items to achieve/identify:

1. Multilateral education efforts to expand awareness of existing detection technologies among those in the content and technology communities
2. Economic and other incentives to make detection technologies more available, accessible and affordable to those lacking in technical knowledge or capital
3. Basic consensus about which technologies are the most reliable and which may require further optimization, along with promising innovations on the horizon; both of which can inform education and development efforts
4. Recommendations to match/merge source file to location request data for undisputed and properly noticed takedowns, with clear and enforceable requirements including properly expressed ownership metadata and experimental dispute resolution protocol(s)
5. Counter-notice mechanisms that are transparent and accessible, and work for a range of stakeholders, from individual creators and founders to large-scale rightsowners/administrators and services
6. A permanent or semi-permanent voluntary standards body that regularly examines the results of voluntary agreements and best practices within the sector and makes recommendations in consultation with stakeholders
7. Best practices for complying with legal directives around fair use with community-defined, non-legal standards within §107 provisions

There are additional areas in which stakeholders could develop novel market solutions within the §512 framework. One idea is of a provisional license for emerging services that have a clear social/UGC component with strict growth/terminate benchmarks and a requirement to make use of community-defined, best class identification, merge and notice tools. Encouraging innovation, economic growth and cross-sector partnerships is possible within the existing §512 framework—it just requires a commitment to thinking creatively in terms of deal structures that contemplate the entirety of the statute, and not just the parts that would serve one interest or the other. Another concept worthy of exploration is a joint licensing project between current §512 adversaries in a new use environment, such as VR (virtual reality) or AU (augmented/ambient reality). Using §512 in a more collaborative fashion to expand market opportunities for developers and content creators/owners would be a major turning point in the evolution of digital distribution and access.

What should be clear from our comments is that both the creative sector and the technology sector would be better served to collaborate rather than invoke conflict at a time when any spark can become a conflagration. Artists, in particular, depend on their industry cohorts to determine the broad frameworks of participation in culture markets. More than being a means to business ends, creators are crucial to the entire ecosystem. Their well-being must be top priority or we will lose more than any reasonable person would care to conjure. Future of Music Coalition offers itself as a resource or partner to any of those who seek a saner, more productive digital content ecology.

Respectfully Submitted,

Casey Rae
CEO
Future of Music Coalition