1)       Cartoon Network, LP v. CSC Holding Inc.

Among the most recent and important decisions impacting cloud-based storage and distribution of entertainment content was the Second Circuit’s 2008 decision in Cartoon Network, LP v. CSC Holding Inc. (“Cablevision Case”), which addressed the copyright implications of a cloud-based DVR system. Specifically, cable operator Cablevision Systems Corporation (“Cablevision”) announced plans in March 2006 to market a “remote storage DVR system” (“RS-DVR”) to allow subscribers without a stand-alone DVR to record cable programs on central hard drives Cablevision maintained at a “remote” location. In response, major networks and studios sued Cablevision in federal court, claiming that the RS-DVR would directly infringe their rights to reproduce and publicly perform their copyrighted works. The district court agreed and enjoined Cablevision from operating the RS-DVR system without additional licenses from the plaintiffs. Cablevision appealed and the Second Circuit reversed the decision on all three infringement counts.

The first claim rejected by the Appeals Court was that the brief caching of buffering data while Cablevision’s system queried whether the customer had actually requested the program be recorded on to the applicable hard drive violated the exclusive right of reproduction. The Appeals Court held that the buffering period was so negligible as to fail the Copyright Act’s requirement that a copy of work be fixed in a tangible medium “for more than a transitory duration.” This point is potentially significant for future cloud business models given that cloud-based services may enable the storage, manipulation and distribution of content in multiple formats across multiple devices, which will continue to no doubt further implicate transitory caching of content at multiple stages in the process.

The second claim reversed by the Appeals Court was that Cablevision was liable for direct copyright infringement for copying programs to the RS-DVRs. Here, the Appeals Court held that Cablevision did not evince the required “volitional conduct” that actually caused the copy to be made and found that Cablevision’s conduct in designing, housing, and maintaining a system did not amount to direct infringement.

The third and perhaps most controversial claim reversed by the Appeals Court was that the transmission of programming from the RS-DVR to subscribers who requested playback breached the public performance right. Here, Cablevision argued (and the Appeals Court found relevant) that, “because each RS-DVR transmission is made using a single unique copy of a work, made by an individual subscriber” only one subscriber is capable of receiving the transmission of that particular work and thus the performance is not “public”.

Ultimately, the holding in the Cablevision Case that individualized copies of content specifically streamed to subscribers from remote DVRs constitute private, as opposed to public, performances introduces a lack of clarity regarding the rights necessary for cloud-based transmissions of audio/visual content. Put simply, the question remains as to whether streaming of legally obtained content to an end user from the cloud (e.g., MP3 tracks stored in a digital locker, etc.) implicates the public performance right. As a result, purveyors of cloud-based business models are left considering whether additional authorization is required from copyright holders, and in the absence of obtaining that consent, whether the potential exists that another tribunal could later disagree with the Second Circuit’s holding in the Cablevision Case. Furthermore, the application of the holding in the Cablevision Case to alternate fact patterns and business models, as even the court itself acknowledged, provides limited guidance.

“This holding, we must emphasize, does not generally permit content delivery networks to avoid all copyright liability by making copies of each item of content and associating one unique copy with each subscriber to the network, or by giving their subscribers the capacity to make their own individual copies. We do not address whether such a network operator would be able to escape any other form of copyright liability, such as liability for unauthorized reproductions or liability for contributory infringement.”

2) Capitol Records, LLC et al. v. MP3tunes, LLC

Another recent and ongoing case potentially impacting the digital locker and cloud computing landscape is Capitol Records, LLC et al. (“EMI”) v. MP3tunes, LLC (See Initial Complaint, EMI Summary Judgment Memorandum and Response, MP3tunes Summary Judgment Memorandum and Response, and recent Summary Judgment Ruling). Here, multiple record companies and publishers affiliated with EMI have asserted, among others, various copyright infringement claims against MP3tunes, which operates two separate online services–specifically, MP3tunes.com and Sideload.com. MP3tunes.com allows users to store their music collections in online digital lockers, which they can then access from any computer or mobile device with an Internet connection. Sideload.com is a music search engine site that allows end users to search for links on the internet to downloadable music that can be uploaded (or “sideloaded”) to an MP3tunes digital locker. Once music is placed in an end user’s digital locker, the music becomes available for transmission to any IP-enabled device at the end user’s direction.

Based on the documents filed to date, EMI has asserted a series of both direct and secondary copyright infringement claims against MP3tunes, including claims that MP3tunes has forfeited its eligibility under the Digital Millennium Copyright Act’s (DMCA) Safe Harbor provisions for its illicit conduct in knowingly providing the means for end users to violate EMI’s copyrights via Sideload.com and failing to respond to takedown notices. In fact, the majority of EMI’s claims are based on the functionality and content made available via Sideload.com, which essentially aggregates URLs linked to digital music files that can be readily downloaded or sideloaded to an online locker. EMI has asserted that the infringing nature of the links posted on Sideload.com, as well as the corresponding files that are made available via the linked URLs that are then sideloaded into an MP3tunes digital locker violate EMI’s copyrights.

While a detailed analysis of the merits of EMI’s DMCA and contributory liability theories remain outside the scope of this post (and have yet to be entirely decided by the court), the district court, in its recent ruling on the parties’ summary judgment motions, did find that the MP3tunes was entitled to the Safe Harbor protections afforded under the DMCA, but further addressed a key issue emerging in the new cloud-based lockering environment. Specifically, upon receipt of a valid takedown notice from EMI, the court found that MP3tunes had a duty to not only remove links to infringing songs publicly displayed on Sideload.com, but also a duty to remove songs stored in users’ personal lockers which were downloaded from such links. In its defense, MP3tunes claimed that it was only required to remove the URL links on Sideload.com because only those links were listed on EMI’s takedown notices and that it might be subject to lawsuits by users if it actually removed personal property from users’ digital lockers. The court, however, rejected this argument, pointing to the DMCA’s immunity provisions for service providers acting on valid takedown notices (see 17 U.S.C. 512(g)), and stating that:

“Where service providers such as MP3tunes allow users to search for copyrighted works posted on the internet and store those works in private accounts, to qualify for DMCA protection, those service providers must (1) keep track of the source and web address of stored copyrighted material, and (2) take content down when copyright owners identify the infringing sources in otherwise compliant notices… [Accordingly,] MP3tunes was obligated to remove specific works traceable to users’ lockers .. [b]ecause MP3tunes keeps track of the source and web address for each sideloaded song in each user’s locker and EMI’s notices gave sufficient information for MP3tunes to locate copies of infringing song in users lockers.

In addition to the foregoing claims, EMI also claimed that MP3tunes directly infringes that right of public performance by allowing end users to stream music from their online digital lockers to personal devices. Relying in part on the holding in the Cablevision Case, EMI asserted that MP3tunes violates the public performance right because it uses a “single master” to play songs to multiple users, as opposed to Cablevision which maintained a separate copy of each program for each subscriber who recorded it. In response, MP3tunes replied that it does not utilize a “single master” storage system, but rather a common open source distributed file software system that eliminates redundancy and enables MP3tunes to efficiently store and retrieve the millions of audio files uploaded by its users without employing a duplicative file storing method.

Ultimately, the district court held that MP3tunes does not in fact use a “single master” system, but rather a standard algorithm known as “Content-Addressable Storage” to store music files which uses hash tags associated with each uploaded song that ultimately allows for the reconstruction of the exact file the user originally uploaded to the service (i.e., there is no “master copy” of any EMI songs stored on MP3tunes’ servers). Still, this determination does not entirely address EMI’s infringement claim regarding the right of public performance as the court’s holding solely relates to the nature of the specific file storage technology employed by MP3tunes. This is in part due to the fact that EMI’s arguments on the public performance issue were largely based on distinguishing the file storing technology used by MP3tunes from the technology employed by Cablevision. In other words, EMI did not address the public performance question by looking at the intended audience of the transmissions enabled by MP3tunes, but rather whether MP3tunes used a “single master” to transmit music to end users.

Ultimately, the take away from both the Cablevision Case and the MP3tunes case is that cloud-based delivery, storage and consumption of entertainment content, whether overtly and implicitly, implicates many of the exclusive rights afforded copyright holders and stakeholders need to remain vigilant about allocating risk when the laws in the US and overseas have yet to suitably address the contours of these services and the corresponding technologies at play. We will obviously keep an eye on future developments in connection with cloudifcation of entertainment content and any case law potentially impacting the future deployment of related cloud-based products and services.

As unique as the service is though, it appears that many of its features were designed and implemented to enable Turntable.fm to operate as a “non-interactive” service under the Digital Millennium Copyright Act (“DMCA”), similar to an internet radio station, thus avoiding the need for direct licenses from the music labels.  For Turntable.fm, the distinction could mean the difference between sustaining a viable business or joining a long line of digital music services that were unable to survive because of the burden of paying license fees to the labels.  While it is too early to determine if this strategy will be challenged and/or whether it will ultimately prevail, Turntable.fm’s service clearly raises some unique legal issues.

To qualify as a “noninteractive” service under the DMCA, a service’s programming cannot be “specially created for the [user]” and cannot “substantially” consist of user-requested songs if those requests are filled within one hour of being made.  On first blush, it would appear that Turntable.fm would have a difficult time meeting these criteria.  After all, one of the core features of the service is the users’ ability to select songs to be played in whatever room the user is visiting, essentially enabling users to be DJs in a virtual music club.

However, Turntable.fm has devised various rules for the service that appear to be tailored directly to address other provisions of the DMCA which determine whether a service can be classified as “noninteractive”.  For example, if there is just one DJ in a room, then the DJ can only hear 30-second previews of the songs selected.  This means that Turntable.fm cannot be used to provide interactive, on-demand programming solely for the user.  In addition, listeners in a room cannot see what song the next DJ plans to play, which satisfies the DMCA requirement that a noninteractive service cannot reveal upcoming songs to be played and while a song is being played, artist and album information is displayed.  Turntable.fm also complies with the requirement that a “noninteractive” service cannot, in any three hour period, play more than three songs from the same album (and no more than two consecutively) and no more than four songs from the same artist (and not more than three consecutively).

Yet, notwithstanding the above restrictions, the DJs have complete control over the music played and it could be argued that this type of user control is unprecedented in a noninteractive service.

The leading case to address the issue of whether a streaming service should be deemed “interactive” or “noninteractive” under the DMCA involved a suit brought by Arista Records against Launch Media, which operates the LAUNCHcast service.  The Second Circuit approached the issue by focusing on the potential impact that a service could have on sales of recorded music.  A “noninteractive” service does not require mechanical and sound recording licenses and thus the additional revenues that the record labels would receive.  Therefore, the court reasoned, at the point that a streaming music service provides “sufficient control to users such that playlists are so predictable that users will choose to listen to the [service] in lieu of purchasing music, thereby—in the aggregate—diminishing record sales,” that service is no longer “noninteractive.”  The court went on to hold that the LAUNCHcast service is “noninteractive” stating that, “to the degree that LAUNCHcast’s playlists are uniquely created for each user, that feature does not ensure predictability.”  In making this finding, the court pointed to the fact that 60% of the songs programmed through the service are done so with virtually no input.  As the court said, “the unique nature of the playlist helps Launch ensure that it does not provide a service so specially created for the user that the user ceases to purchase music.”

Whether Turntable.fm can meet the “predictability” test established by the Launch Media decision or whether it would otherwise be deemed to comply with the DMCA’s requirements for a “noninteractive” service is open for debate.  A reasonable argument could be made that users of the service have substantial control over the songs being played, making the programming sufficiently predictable that Turntable.fm could become a substitute for purchasing music.

Beyond the Second Circuit’s analysis under the Launch Media decision, other features of the Turntable.fm service do not appear to comply with certain provisions of the DMCA’s definition of a “noninteractive” service.  The music is programmed entirely by the user DJs and streamed within one hour of “selection” in violation § 114(j)(7) of the DMCA.

More broadly, the DMCA defines an “interactive” service as one that is “specially created” for the user.  It is unclear how the “specially created” element might be interpreted for a service like Turntable.fm, which has both DJ users and “audience” users.  Depending on the vantage point (i.e. DJ user or “audience member” user), the “specially created” factor could come out on either side of the “interactive”/”noninteractive” divide.

Of course, unless and until Turntable.fm is challenged in court, these questions may never be put to the test.  But regardless of whether they are, we do not believe that the unanswered questions will slow the rapid evolution of the digital music landscape and anticipate the continued launch of new services in the coming months and years.

**  Justin Greenbaum, a summer associate with the Firm, assisted in the preparation of this piece.

The suit, originally filed in 2001 by several labels owned by Sony BMG, including Arista, Bad Boy and Zomba, alleged that LAUNCHcast, a webcasting service run by Yahoo’s Launch Media unit, which enables users to create “stations” that play songs within a particular genre or similar to a particular artist or song, violated provisions of the DMCA that required payment of licensing fees for the use of sound recordings in an “interactive” service.” 

The DMCA defines an interactive service as one “that enables a member of the public to receive a transmission of a program specially created for the recipient, or on request, a transmission of a particular sound recording . . . which is selected by or on behalf of the recipient.”   The provision was intended to distinguish individual playbacks from transmissions “whose primary purposes . . . is to provide to the public such audio or other entertainment programming.” 

That distinction was critical.  The DMCA requires payment of royalties, based on a negotiated license, to the owner of a sound recording (in most instances, a record label) for use in an interactive service.  For non-interactive services, webcasters need only pay a compulsory or statutory licensing fees set by the Copyright Royalty Board, which are considerably lower than individual, negotiated license fees. 

The rationale behind the differing licensing schemes was to protect the recording industry from lost sales.  As the Second Circuit stated “If the user has sufficient control over the interactive service such that she can predict the songs she will hear, much as she would if she owned the music herself and could play each song at will, she would have no need to purchase the music she wishes to hear.” 

In affirming the jury’s finding that LAUNCHcast’s service is not “interactive” under the DMCA, the Second Circuit analyzed the functions and features of the service itself, focusing on whether a user can receive a transmission of a program “specially created” for him or her.  For the Court, the key issue was “predictability”–to fall within the provisions requiring payment of a sound recording license fee, a service must provide “a degree of predictability-based on choices made by the user-that approximates the predictability the music listener seeks when purchasing music.”  The appellate panel found that the LAUNCHcast service does not provide sufficient control to users such that the playlists selected are so predictable that users will choose to listen to the service instead of purchasing music. 

The court reached that conclusion based on three primary factors.  First, the rules governing what songs are pooled in a playlist ensure that the user has no ability to control–much less predict–which songs are pooled for selection.  The panel noted that at least 60% of the songs are generated by factors entirely beyond the user’s control. 

Second, the service’s rules prevent a user’s explicitly rated songs–presumably the ones he or she wants to listen to–from being anywhere near a majority of the songs on a playlist.  At a minimum, 20% of the songs are unrated.  In addition, when a user rates a particular song, the service “implicitly” rates all the other songs by that artist, subjecting the user to many songs the user may never have heard or may not like.

Lastly, the court noted that LAUNCHcast randomly orders playlists–taking into account the DMCA’s restrictions on the consecutive play of artists or albums–further restricting the user’s ability to choose artists or albums they wish to hear.

The court concluded that the only thing a user can predict with certainty is that by rating a song at “zero” the user will not hear that song on that station again.

The decision is significant for several reasons.  It is the first time a federal appeals court has undertaken a careful, detailed and well-reasoned analysis of the DMCA’s provisions related to sound recording performance rights.  The decision clearly sets forth the rationale for the different licensing schemes for web transmission of music established by the DMCA, removing a considerable amount of uncertainty to webcasters and others who feared that their business models might be in jeopardy if they faced additional licensing fees.  And it parsed the features and functions of the LAUNCHcast service.  In doing so, the court has provided significant guidance on the legal, business and technical fronts which may enable future growth and innovation in the space.

As we discussed in a previous post, Lenz, in her suit against Universal, is asserting a claim of “misrepresentation” under Section 512(f) of the DMCA, which states that

“Any person who knowingly materially misrepresents under this section that material or activity is infringing…shall be liable for any damages, including costs and attorneys’ fees, incurred by the alleged infringer… who is injured by such misrepresentation, as the result of the service provider relying upon such misrepresentation in removing or disabling access to the material or activity claimed to be infringing…”

Lenz’s misrepresentation claim, therefore, could be sustained only if Universal failed to comply with the takedown requirements under Section 512(c)(3)(A)(v) which states that any copyright holder filing a takedown notice must have “a good faith belief that use of the material in the manner complained of is not authorized by the copyright owner, its agent, or the law.”

In response to Lenz’s suit, Universal filed a motion to dismiss claiming that copyright owners cannot be required to evaluate the question of fair use prior to sending a takedown notice because fair use is merely an excused infringement of copyright rather than a use authorized by the copyright holder or by law. In other words, the “good faith” requirement under Section 512(c)(3)(A)(v) does not apply because fair use is not a use authorized by law. Lenz (supported by lawyers from the Electronic Frontier Foundation), countered that fair use is, in fact, an authorized use of copyrighted material and that copyright holders cannot represent in good faith under the DMCA that content infringes copyright as required without considering all authorized uses of such content, including fair use.

In August, the U.S. District Court for the Northern District of California, rejected Universal’s motion to dismiss and held that the Copyright Act provides explicitly that “the fair use of a copyrighted work…is not an infringement of copyright” and that fair use is lawful use of copyright. Accordingly, the court held that Section 512(c)(3)(A)(v) applies to fair use of a copyright and therefore Universal could be held liable on a claim of misrepresentation if Lenz can prove the requisite standard of subjective bad faith necessary to prevail.

While the impact of this decision is not yet clear (as it is merely a dismissal of a motion that allows the case to proceed), copyright holders need to proceed with caution in their approach to issuing takedown notices. If the district court’s initial ruling is upheld, it is possible that copyright holders may be required to consider the potential fair use applications of a copyrighted work before issuing takedown notices and the failure to do so may constitute a violation of the DMCA. In other words, corporations and their intellectual property enforcement regimes will now be tasked with making value judgments on the applicability of the fair use doctrine—one of the most amorphous and continually evolving legal concepts in existence.  

The DMCA was initially intended as the US implementation of two treaties adopted by the World Intellectual Property Organization (WIPO) in 1996 to establish rules for two evolving forms of digital media: music and computer software and databases.  However, as the legislation was introduced in Congress, additional provisions were added in response to lobbying efforts by two distinct constituencies.  

Media companies wanted the bill to include provisions to protect their IP, which was increasingly available in digital form, from widespread infringement.  They therefore pressed for the inclusion of anti-circumvention rules that prevent anyone from bypassing any technological measures used by copyright owners to control access to their works.  And ISPs, hosting companies and interactive services sought provisions to provide a safe harbor from infringement claims based on the actions of their users.  In hindsight, these provisions, which were ultimately included in the DMCA, have each impacted the technological landscape in ways that are felt everyday

Without the anti-circumvention provisions, it is unlikely that DVD technology, which was being test marketed when the DMCA was enacted, would have been embraced by movie studios.  The adoption of DVD technology has led to the blossoming of new lines of electronics manufacturing and rental businesses that have far outpaced those which evolved when VHS was the primary medium for home viewing of movies.

However, these same provisions have also led to disputes and litigation that, to some, run counter to their original intent.  For example, while developing and improving methods to ensure the security of the internet and computers connected to it would seem to be an unassailable goal, the anti-circumvention provisions have created a Catch-22 for security experts and researchers.  In order to improve copy-protection systems and computer security programs, these experts and researchers need to determine what flaws exist in any currently deployed security system.  Thus, the very act of testing those systems is a violation of the anti-circumvention provisions of the DMCA.  For a more detailed, albeit one-sided, view of this debate, here is the Electronic Frontier Foundation’s report, entitled “Unintended Consequences: Ten Years under the DMCA.”

The safe harbor provisions have enabled the explosive growth of the web, particularly as an interactive medium.  In effect, website operators from MySpace to WordPress were able to provide forums for users without a constant fear of being sued based on the acts of their users.

In order to maintain “safe harbor” protection, the “provider of online services” must meet certain conditions.  It cannot have knowledge of an infringement nor derive any direct financial benefit from the infringing activity where it has the right and ability to control it.  A provider must also expeditiously remove material if a copyright holder sends a take-down notice, without the need to evaluate the notice to determine if it is accurate or whether the user who uploaded the content has any rights to it.   From a simple perspective, this seems to make sense:  a provider should not be put in a position of having to adjudicate disputes over content.  As a pure intermediary, its obligations should be as simple as possible (If receive take-down notice è then take down content).  The principals are the ones with the vested interest in the content and should be the only ones debating the merits of who has the right to make content available.

However, some content owners believe that too much emphasis is being placed on the take-down “remedy” enabling it to serve as a blanket of immunity for all infringement claims.  What if an online provider’s sole business model is to provide users with a readily available means to post content online?  And what if the site becomes wildly popular, with the potential to earn substantial amounts of money selling ads to companies seeking that vast user base and audience? And what if it is clear that users are posting content they don’t own?  Doesn’t the online provider lose its safe harbor protection because it has knowledge of the infringement?  Or is deriving financial benefit from the infringement? 

These are the basic facts underlying Viacom’s lawsuit against Google.  The case is headed for trial next spring.  The outcome will undoubtedly have major repercussions throughout the digital media world.

The intermediary function of online providers has also recently come under assault in a somewhat surprising context, the presidential campaign.  John McCain’s campaign had posted various videos on YouTube that included snippets from broadcast news footage.  In response to take-down notices from the news outlets, YouTube pulled the videos.

The McCain campaign then sent a letter to YouTube claiming its use of the footage was privileged under the fair use doctrine.  While recognizing that a fair use analysis could not be undertaken in response to every take-down notice, the McCain campaign proposed that such an analysis be done, at the very least, when the content at issue was posted by a political campaign.

In response, YouTube, politely, told the McCain campaign “thanks but no thanks.”  In a possible overstatement, the letter said that “[l]awyers and judges constantly disagree about what does and does not constitute fair use.  No number of lawyers could possibly determine with a reasonable level of certainty whether all videos for which we receive disputed takedown notices qualify as fair use.”  While acknowledging that some parties abuse the take-down notification process, YouTube essentially passed the burden of policing those abuses to the users who upload the content who, according to YouTube, “operat[e] from a position of strength, with knowledge of exactly where the content in [their] videos came from.”

Ironically, McCain voted for the DMCA.  But YouTube’s response appears to be more in line with both the express language and intent of the safe harbor and take-down notice provisions of the act.

The question of whether a fair use analysis needs to be employed in connection with issuing a take-down notice is currently the subject of litigation involving not an online provider but content owners.  Universal Music and Universal Publishing Group had issued a take-down notice to YouTube over a 29-second video posted by Stephanie Lenz of her toddler dancing to Prince’s “Let’s Go Crazy.”  YouTube took down the video.  Lenz then sued the Universal entities claiming that they made a misrepresentation in sending the take-down notice to YouTube knowing that Lenz’ use of “Let’s Go Crazy” was not infringing.  In essence, Lenz was arguing that a copyright owner must make a fair use analysis prior to sending a take-down notice to an online provider.

The Universal entities moved to dismiss the complaint.  At issue is language in the DMCA that, when sending a take-down notice, the complaining party must state its good faith belief that the use at issue is “not authorized by the copyright owner.”  Universal argues that fair use is merely an “excused” infringement and therefore a fair use evaluation is not relevant to determine whether the use is “not authorized” for DMCA purposes.  As Universal also pointed out, the DMCA take-down provisions do not even mention fair use, much less place a burden on the party issuing the notice to consider it.

The district court denied the motion, stating that the issue of whether fair use qualifies as “authorized by law” in connection with a take-down notice under the DMCA was a question of first impression. 

Obviously, this is a case that will be closely watched.  However, a final ruling in Lenz’ favor will have considerable consequences, effectively shifting the fair use burden from the user uploading the content (as YouTube noted in its letter to the McCain campaign) to the party issuing the take-down notice, in most cases copyright right owner. 

And so, we usher in the next decade of the DMCA.  And while it is difficult to predict what the next ten years will bring, it is probably safe to say that, regardless of the disputes that will inevitably arise, the digital marketplace will continue to rapidly evolve, with the DMCA serving as the basic framework and foundation for protecting digital rights and the focal point for disputes yet to come.

    Winning the Digital Race

In fact, in its attempt to prevent future instances of what the AP is claiming to be copyright infringement by bloggers, it has made public its intention to develop guidelines for what would be considered permissible use (e.g., a “fair use”) of AP’s content. Is it just me, or does the all-too-familiar law school term “slippery slope” come to mind? Is the AP essentially trying to create its own legal definition of what qualifies as “fair use”? The AP says no, but many bloggers, and I’m sure many intellectual property attorneys alike, are skeptical. Although Cadenhead’s attorney seems to think that it might be helpful to have insight into what AP considers to be acceptable use of their content in blogs (http://www.msnbc.msn.com/id/25329749/from/ET/), courts traditionally make fair use determinations on a case-by-case basis, depending on a variety of fact-specific circumstances surrounding the way in which a particular piece of copyrighted content is used. Therefore, the AP’s guidelines could be construed as an encroachment upon the legal process of evaluating arguments of fair use.