Court of Appeals Determines that Launchcast is Not an Interactive Service - Thus Not Needing Direct Licenses From the Record Labels
The question of when a digital music service is “interactive” and therefore requires direct negotiations with a copyright holder in order to secure permission to use a sound recording is a difficult one that has been debated since the Digital Millennium Copyright Act was adopted in 1998. In a decision of the Second Circuit Court of Appeals released today, upholding a jury decision in 2007, the Court concluded that Yahoo’s Launchcast service (now operated by CBS) is not so “interactive” as to take it outside of the statutory royalty despite the fact that the service does customize its music offerings to the tastes of individual listeners. To reach its decision, the Court went through an extensive analysis of both the history of the sound recording copyright and of the details of the criteria used by Launchcast to select music for a stream sent to a specific user. By determining that the service is not interactive, the service need only pay the SoundExchange statutory royalty to secure permission to use all legally recorded and publicly released music. Had the service been found to be interactive within the meaning of the statute, the service would have to negotiate with each sound recording copyright holder for each and every song that it wanted to use on its service to get specific rights to use each song - potentially resulting in hundreds of negotiations and undoubtedly higher fees than those paid under the statutory license.
The issue in the case turned on an analysis of the DMCA’s definition of an interactive service. The statute defines an interactive service as one where a user can select a specific song or “receive a transmission of a program specially created for the recipient.” It is clear that Launchcast did not allow a user to request and hear a specific song. But, by specifying a genre of music, and by specifying favorite artists and songs and rating other songs played by the service, a listener could influence the music that was provided to it. Was this ability to influence the music sufficient to make it an “interactive service” and thus take it out of the coverage of the statutory royalty?
After an exhaustive analysis of the process that Launchcast goes through to create a stream for a listener, the Court focused on several facts. First, the Court found that much of the music in any stream delivered by Launchcast was not music selected by the user in their list of preferred artists and songs, but was instead picked by Launchcast from its vast library of songs using a number of factors. The Court also made clear that listeners had no ability to game the system to make it play more favorites of the listener. While a listener could skip some songs, and pause a song in the middle of its play, it could not go backwards to replay songs or otherwise make particular songs play more frequently. In short, the Court found that the system was set up so that it would not substitute for the purchase of music as listeners could not get songs or even particular artists when they wanted. The Court used the term “predictability” – and found that the user had no predictability in determining whether or when any specific song would play during any listening session, and thus the service was not a substitute for a purchase of a song.
This was important in the Court’s analysis. First, the Court determined that the phrase in the statute defining an interactive stream to be “a transmission specially created for the recipient” was not a model of clarity, and was capable of many interpretations. While the record companies argued that any stream that was created specifically for a user based on the user’s preferences was, by definition, “specially created for the recipient”, the Court found that such a simplistic view could not be sustained. Instead, the language of the statute has to be interpreted in light of the intent of Congress in the adoption of the statute. The Court went through a thorough analysis of the history of the sound recording royalty and how the DMCA provision at issue here came to be in the 1998 Act. The Court noted that the sound recording performance right was first adopted in the US in 1995 and was intended to be a narrow right, initially being applied only to subscription services. After its adoption, upon fears of piracy on the Internet, the right was expanded three years later to include noninteractive streams. In enacting the broader performance royalty, the DMCA broadened the definition of an interactive stream to include the phrase at issue here, focusing primarily on the issue of digital piracy and the fear that a predictable stream of music would allow digital copying. The Court cites specific language of the House of Representatives report on the DMCA where the House stated that you have an interactive stream “if a transmission recipient is permitted to select particular sound recordings in a prerecorded or predetermined program.”
After looking at the history and the way the service functioned, the Court focused on the language of the statute that said that there had to be a “transmission of a program” that was specially created for the user before the program was deemed to be interactive. The definition of a "transmission of a program" looked at the transmission of a program as a whole – to find that there was an interactive transmission of a program one has to look at the entire transmission to see if the entire transmission was created specially for the user. The Court determined that, given the way the Launchcast system was set up, the user was really able to specifically influence only a small number of songs that were played in his or her stream. The vast majority of the songs were selected by Launchcast and would be of the same genre as the listener's preferences, but what the songs would be was not at all predictable. Finding that the user thus had no predictability in the entirety of the program that was transmitted, the Court found that the streams would not significantly substitute for the purchase of specific music, and thus should not be considered interactive in the meaning that Congress intended.
The decision is very interesting in the depth of its specific analysis of the methodology for the formation of a playlist by Launchcast, and in its examples of music references sprinkled throughout (references to U2's Joshua Tree CD, to Gordon Lightfoot and the Beatles, and to “‘special requests’ [on AM radio which] represented love-struck adolescents’ attempts to communicate their feelings to that ‘special friend.’” The judges also candidly acknowledge that they are “appointed for life” and thus have varying degrees of familiarity with the technology which they are discussing.
What is the impact of the case? It undoubtedly helps solidify the position long taken by webcasters that some degree of user influence is permissible by a service that relies on the statutory license for noninteractive webcasting. However, the decision was very fact dependent, with few clear boundaries as to what percentage of a stream can be user influenced and what degree that influence can be exercised to remain within the statutory license. Moreover, this is the decision of a single Court of Appeals – albeit an important one sitting in New York, covering the Northeast, and very active on copyright issues. But other cases in other circuits would not be bound by this decision, though they will no doubt find it to be instructive. But, with other facts, any court might not reach the same decision. Thus, the question of which streams are interactive requiring that a service get a negotiated license from each copyright holder to perform the sound recordings, and which are noninteractive and can be streamed simply by paying SoundExchange the statutory royalty (and I say “simply” with a grain of salt given the multiplicity of options for paying the statutory royalty), will no doubt not be put to rest by this one decision.
