Tag: actual knowledge

Categories
Copyright

Scalia Law Students and CPIP Scholars Make an Impact in Copyright Office Section 512 Study

the word "copyright" written on a typewriterThe U.S. Copyright Office released its long-awaited report on Section 512 of Title 17 late last week. The Report is the culmination of more than four years of study by the Office of the safe harbor provisions for online service provider (OSP) liability in the Digital Millennium Copyright Act of 1998 (DMCA). Fortuitously, the study period coincided with the launch of Scalia Law’s Arts and Entertainment Advocacy Clinic. Clinic students were able to participate in all phases of the study, including filing comments on behalf of artists and CPIP scholars, testifying at roundtable proceedings on both coasts, and conducting a study of how OSPs respond to takedown notices filed on behalf of different types of artists. The Office cites the filings and comments of Scalia Law students numerous times and ultimately adopts the legal interpretation of the law advocated by the CPIP scholars.

The Office began the study in December 2015 by publishing a notice of inquiry in the Federal Register seeking public input on the impact and effectiveness of the safe harbor provisions in Section 512. Citing testimony by CPIP’s Sean O’Connor to the House Judiciary Committee that the notice-and-takedown system is unsustainable given the millions of takedown notices sent each month, the Office launched a multi-pronged inquiry to determine whether Section 512 was operating as intended by Congress.

Scalia Law’s Arts and Entertainment Advocacy Clinic drafted two sets of comments in response to this initial inquiry. Terrica Carrington and Rebecca Cusey submitted comments to the Office on behalf of middle class artists and advocates, including Blake Morgan, Yunghi Kim, Ellen Seidler, David Newhoff, and William Buckley, arguing that the notice-and-takedown regime under Section 512 is “ineffective, inefficient, and unfairly burdensome on artists.” The students pointed out that middle class artists encounter intimidation and personal danger when reporting infringements to OSPs. Artists filing takedown notices must include personal information, such as their name, address, and telephone number, which is provided to the alleged infringer or otherwise made public. Artists often experience harassment and retaliation for sending notices. The artists, by contrast, obtain no information about the identity of the alleged infringer from the OSP. The Office’s Report cited these problems as a detriment for middle class artists and “a major motivator” of its study.

A second response to the notice of inquiry was filed by a group of CPIP scholars, including Sandra Aistars, Matthew Barblan, Devlin Hartline, Kevin Madigan, Adam Mossoff, Sean O’Connor, Eric Priest, and Mark Schultz. These comments focused solely on the issue of how judicial interpretations of the “actual” and “red flag” knowledge standards affect Section 512. The scholars urged that the courts have interpreted the red flag knowledge standard incorrectly, thus disrupting the incentives that Congress intended for copyright owners and OSPs to detect and deal with online infringement. Several courts have interpreted red flag knowledge to require specific knowledge of particular infringing activity; however, the scholars argued that Congress intended for obvious indicia of general infringing activity to suffice.

The Office closely analyzed and ultimately adopted the scholars’ red flag knowledge argument in the Report:

Public comments submitted by a group of copyright law scholars in the Study make a point closely related to the rightsholders’ argument above, focusing on the different language Congress chose for actual and red flag knowledge. They note that the statute’s standard for actual knowledge is met when the OSP has “knowledge that the material or an activity using the material on the system or network is infringing” or “knowledge that the material or activity is infringing,” while the red flag knowledge standard is met when the OSP is “aware of facts or circumstances from which infringing activity is apparent.” This difference, the copyright law scholars argue, is crucial to understanding the two standards: while the statute uses a definite article—“the”—to refer to material or activity that would provide actual knowledge, it drops “the” to speak more generally about facts or circumstances that would create red flag knowledge. “In Congress’s view,” the comment concludes, “the critical distinction between the two knowledge standards was this: Actual knowledge turns on specifics, while red flag knowledge turns on generalities.”

 

The Office went on to state that “a standard that requires an OSP to have knowledge of a specific infringement in order to be charged with red flag knowledge has created outcomes that Congress likely did not anticipate.” And since “courts have set too high a bar for red flag knowledge,” the Office concluded, Congress’ intent for OSPs to act upon information of infringement has been subverted. This echoed the scholars’ conclusion that the courts have disrupted the balance of responsibilities that Congress sought to create with Section 512 by narrowly interpreting the red flag knowledge standard.

Scalia Law students and CPIP scholars likewise participated in roundtable hearings on each coast to provide further input for the Copyright Office’s study of Section 512. The first roundtable was held on May 2-3, 2016, in New York, New York, at the Thurgood Marshall United States Courthouse, where the Second Circuit and Southern District of New York hear cases. The roundtable was attended by CPIP’s Sandra Aistars and Matthew Barblan. They discussed the notice-and-takedown process, the scope and impact of the safe harbors, and the future of Section 512. The second roundtable was held in San Francisco, California, at the James R. Browning Courthouse, where the Ninth Circuit hears cases. Scalia Law student Rebecca Cusey joined CPIP’s Sean O’Connor and Devlin Hartline to discuss the notice-and-takedown process, applicable legal standards, the scope and impact of the safe harbors, voluntary measures and industry agreements, and the future of Section 512. Several of the comments made by the CPIP scholars at the roundtables ended up in the Office’s Report.

In November 2016, the Office published another notice of inquiry in the Federal Register seeking additional comments on the impact and effectiveness of Section 512. The notice itself included citations to the comments submitted by Scalia Law students and the comments of the CPIP scholars. Under the guidance of Prof. Aistars, the students from Scalia Law’s Arts and Entertainment Advocacy Clinic again filed comments with the Office. Clinic students Rebecca Cusey, Stephanie Semler, Patricia Udhnani, Rebecca Eubank, Tyler Del Rosario, Mandi Hart, and Alexander Summerton all contributed to the comments, which discussed their work in helping individuals and small businesses enforce their copyright claims by submitting takedown notices pursuant to Section 512. The students reported on the practical barriers to the effective use of the notice-and-takedown process at particular OSPs. Two problems identified by the students were cited by the Copyright Office as examples of how OSPs make it unnecessarily difficult to submit a takedown notice. Accordingly, the Office called on Congress to update the relevant provisions of Section 512.

Two years after the additional written comments were submitted, the Office announced a third and final roundtable to be held on April 8, 2019, at the Library of Congress in Washington, D.C. The purpose of this meeting was to discuss any relevant domestic or international developments that had occurred during the two prior years. CPIP’s Devlin Hartline attended this third roundtable to discuss recent case law related to Section 512, thus ensuring that CPIP scholars were represented at all three of the Office’s roundtables.

CPIP congratulates and thanks the students of Scalia Law’s Arts and Entertainment Advocacy Clinic for their skillful advocacy on behalf of artists who otherwise would not be heard in these debates.

Categories
Copyright

Twenty Years Later, DMCA More Broken Than Ever

a lightbulb shatteringWith Section 512 of the DMCA, Congress sought to “preserve[] strong incentives for service providers and copyright owners to cooperate to detect and deal with copyright infringements that take place in the digital networked environment.”[1] Given the symbiotic relationship between copyright owners and service providers, Congress meant to establish an online ecosystem where both would take on the benefits and burdens of policing copyright infringement. This shared-responsibility approach was codified in the Section 512 safe harbors. But rather than service providers and copyright owners working together to prevent online piracy, Section 512 has turned into a notice-and-takedown regime where copyright owners do most of the work. This is not what Congress intended, and the main culprit is how the courts have misinterpreted Section 512’s red flag knowledge standards.

Several provisions in Section 512 demonstrate that Congress expected service providers to also play a role in preventing copyright infringement by doing some of the work in finding and removing infringing material. In order to benefit from the safe harbors, service providers must designate an agent to receive takedown notices, respond expeditiously to takedown notices, act upon representative lists, implement reasonable repeat infringer policies, and accommodate standard technical measures. But it’s also clear that service providers have the duty to remove infringing content even without input from copyright owners. As the Senate Report notes, “Section 512 does not require use of the notice and take-down procedure.”[2] And the knowledge provisions in Section 512 reflect this. Absent a takedown notice, service providers must, “upon obtaining . . . knowledge or awareness” of infringing material or activity, “act[] expeditiously to remove . . . the material” in order to maintain safe harbor protection.[3]

Under Section 512, there are two kinds of knowledge that trigger the removal obligation without input from the copyright owner—actual and red flag. For sites hosting user-uploaded content, actual knowledge is “knowledge that the material or an activity using the material on the system or network is infringing”[4] and red flag knowledge is “aware[ness] of facts or circumstances from which infringing activity is apparent.”[5] Thus, actual knowledge requires knowledge that specific material (“the material”) or activity using that specific material (“activity using the material”) is actually infringing (“is infringing”), while red flag knowledge requires only general awareness (“aware[ness] of facts or circumstances”) that activity appears to be infringing (“is apparent”). There are similar knowledge provisions for search engines.[6]

Importantly, actual knowledge refers to “the material,” and red flag knowledge does not. It instead refers to “infringing activity”—and it’s not the infringing activity, but infringing activity generally.[7] However, with both actual and red flag knowledge, the service provider is obligated to remove “the material.”[8] With actual knowledge, the service provider doesn’t have to go looking for “the material” since it already has actual, subjective knowledge of it. But what about with red flag knowledge where the service provider only knows of “infringing activity” generally? How does it know “the material” to take down? The answer is simple: Once the service provider is subjectively aware of facts or circumstances from which infringing activity would be objectively apparent[9]—that is, once it has red flag knowledge—it has to investigate and find “the material” to remove.[10]

The examples given in the legislative history, which relate to search engines, show that red flag knowledge puts the burden on the service provider. As the Senate Report notes, merely viewing “one or more well known photographs of a celebrity at a site devoted to that person” would not hoist the red flag since the images might be licensed or fair use.[11] However, sites that are “obviously infringing because they . . . use words such as ‘pirate,’ ‘bootleg,’ or slang terms . . . to make their illegal purpose obvious . . . from even a brief and casual viewing” do raise the red flag.[12] And a search engine “that views such a site and then establishes a link to it . . . must do so without the benefit of a safe harbor.”[13] Thus, Congress didn’t want search engines worrying about questionable infringements on a small scale, but it also didn’t want search engines to catalog sites that are clearly dedicated to piracy. And, most importantly, red flag knowledge kicks in once a service provider looks at something that is “obviously pirate”—even if it’s an entire website.[14]

So how has Congress’s commonsensical plan worked out? Not very well. As of today, Google indexes nearly 1.5 million results from the infamous pirate site, The Pirate Bay.[15] Like the examples in the legislative history, The Pirate Bay has the word “pirate” in its title, and a brief viewing of the site reveals its obvious, infringing purpose. And it’s not like Google hasn’t been told that The Pirate Bay is dedicated to infringement—nor that it needs to be told. According to data from the Google Transparency Report,[16] the search engine has received requests to remove over 4 million URLs from thepiratebay.org domain alone. There are many other related domains, such as thepiratebay.se, that have received millions of requests as well. In fact, Google tells us that it has received requests to remove over 4 billion URLs from its search engine due to copyright infringement, with many domains receiving more than 10 million requests each.

So how is it that Google can index The Pirate Bay and not be worried about losing its safe harbor? The answer is that the courts have construed Section 512 in a way that contradicts the statutory text and Congress’s intent. They’ve all but read red flag knowledge out of Section 512 and placed the burden of policing infringement disproportionately on the copyright owner. And by narrowing the applicability of red flag knowledge, the courts have perversely incentivized service providers to do as little as possible to prevent infringements. Instead of looking into infringing activity of which they are subjectively aware, they are better off doing nothing lest they gain actual, specific knowledge that would remove their safe harbor protection.

A brief traverse through the case law, especially in the Second and Ninth Circuits, shows how the red flag knowledge train has been derailed. In Perfect 10 v. CCBill, the Ninth Circuit held that domains such as illegal.net and stolencelebritypics.com—the very sort of indicia mentioned in the legislative history—were not enough to raise the red flag.[17] According to the court, “describing photographs as ‘illegal’ or ‘stolen’ may be an attempt to increase their salacious appeal, rather than an admission that the photographs are actually illegal or stolen.”[18] While that’s certainly possible, it’s not likely. And it defies common sense. The Ninth Circuit concluded: “We do not place the burden of determining whether photographs are actually illegal on a service provider.”[19] But this misses the point, and it conflates red flag knowledge with actual knowledge.

The Second Circuit in Viacom v. YouTube held that the “difference between actual and red flag knowledge is . . . not between specific and generalized knowledge, but instead between a subjective and an objective standard.”[20] The court arrived there by focusing on the removal obligation, reasoning that “expeditious removal is possible only if the service provider knows with particularity which items to remove.”[21] And it rejected an “amorphous obligation” to investigate “in response to a generalized awareness of infringement.”[22] By limiting red flag knowledge to specific instances of infringement, the Second Circuit severely curtailed the obligations of service providers to police infringements on their systems. The entire point of red flag knowledge is to place a burden on the service provider to investigate the infringing activity further so that the specific material can be removed.[23]

The Ninth Circuit followed suit in UMG Recordings v. Shelter Capital, holding that red flag knowledge requires specificity.[24] The court reasoned that requiring “specific knowledge of particular infringing activity makes good sense” because it will “foster cooperation” between copyright owners and service providers “in dealing with infringement” online.[25] This “cooperation,” according to the Ninth Circuit, would come from takedown notices sent by copyright owners, who “know precisely what materials they own, and are thus better able to efficiently identify” infringing materials than service providers, “who cannot readily ascertain what material is copyrighted and what is not.”[26] Of course, this is not the shared-responsibility approach envisioned by Congress, and it conflates the red flag knowledge standards with the obligation to respond to takedown notices—separate provisions in Section 512.

Perhaps the most significant gutting of red flag knowledge can be found in the Second Circuit’s opinion in Capitol Records v. Vimeo.[27] The district court below had held that it was a question for the jury whether full-length music videos of current, famous songs that had been viewed by the service provider amounted to red flag knowledge.[28] But the Second Circuit disagreed that there was any jury question: “[T]he mere fact that a video contains all or substantially all of a piece of recognizable, or even famous, copyrighted music and was to some extent viewed . . . would be insufficient (without more) to sustain the copyright owner’s burden of showing red flag knowledge.”[29] That this gloss on red flag knowledge “reduces it to a very small category” was of “no significance,” the Second Circuit reasoned, since “the purpose of § 512(c) was to give service providers immunity, in exchange for augmenting the arsenal of copyright owners by creating the notice-and-takedown mechanism.”[30]

The Second Circuit thus held as a matter of law that there was no need for the factfinder to determine whether the material was so obviously infringing that it would raise the red flag. How is this a question of law and not fact? The court never explains. More importantly, this flies in the face of what Congress provided for with red flag knowledge, and it demotes Section 512 to being merely a notice-and-takedown regime where copyright owners are burdened with identifying infringements on URL-by-URL basis. Giving service providers a free pass when confronted with a red flag turns the Section 512 framework on its head. And it enables service providers to game the system and build business models on widespread, infringing content—even if they welcome it—so long as they respond to takedown notices. The end result is that an overwhelming amount of obvious infringements goes unchecked, and there’s essentially no cooperation between service providers and copyright owners as Congress intended.

[1] S. Rep. No. 105-190, at 40 (emphasis added); see also H.R. Rep. No. 105-551(II), at 49.

[2] Id. at 45; see also H.R. Rep. No. 105-551(II), at 54.

[3] 17 U.S.C. § 512(c)(1)(A)(iii); see also 17 U.S.C. § 512(d)(1)(C).

[4] Id. at § 512(c)(1)(A)(i) (emphasis added).

[5] Id. at § 512(c)(1)(A)(ii) (emphasis added).

[6] See id. at §§ 512(d)(1)(A)-(B).

[7] See 4-12B Nimmer on Copyright § 12B.04[A][1][b][ii] (“By contrast, to show that a ‘red flag’ disqualifies defendant from the safe harbor, the copyright owner must simply show that ‘infringing activity’ is apparent—pointedly, not ‘the infringing activity’ alleged in the complaint.” (emphasis in original)).

[8] 17 U.S.C. § 512(c)(1)(A)(iii)

[9] S. Rep. No. 105-190, at 44 (“The ‘red flag’ test has both a subjective and an objective element. In determining whether the service provider was aware of a ‘red flag,’ the subjective awareness of the service provider of the facts or circumstances in question must be determined. However, in deciding whether those facts or circumstances constitute a ‘red flag’–in other words, whether infringing activity would have been apparent to a reasonable person operating under the same or similar circumstances–an objective standard should be used.”); see also H.R. Rep. No. 105-551(II), at 53.

[10] Id at 48 (“Under this standard, a service provider would have no obligation to seek out copyright infringement, but it would not qualify for the safe harbor if it had turned a blind eye to ‘red flags’ of obvious infringement.”); see also H.R. Rep. No. 105-551(II), at 57.

[11] Id.; see also H.R. Rep. No. 105-551(II), at 57-58.

[12] Id.; see also H.R. Rep. No. 105-551(II), at 58.

[13] Id. at 48-49; see also H.R. Rep. No. 105-551(II), at 58.

[14] Id. at 49; see also H.R. Rep. No. 105-551(II), at 58.

[15] See The Pirate Bay, available at https://www.thepiratebay.org/.

[16] See Google Transparency Report, available at https://transparencyreport.google.com/.

[17] See Perfect 10, Inc. v. CCBill LLC, 488 F.3d 1102 (9th Cir. 2007).

[18] Id. at 1114.

[19] Id.

[20] Viacom Int’l, Inc. v. YouTube, Inc., 676 F.3d 19, 31 (2d Cir. 2012).

[21] Id. at 30.

[22] Id. at 30-31.

[23] See, e.g., H.R. Rep. No. 105-551(I), at 26 (“Once one becomes aware of such information, however, one may have an obligation to check further.”).

[24] UMG Recordings, Inc. v. Shelter Capital Partners LLC, 718 F.3d 1006 (9th Cir. 2013).

[25] Id. at 1021-22.

[26] Id. at 1022.

[27] Capitol Records, LLC v. Vimeo, LLC, 826 F.3d 78 (2d Cir. 2016).

[28] Capitol Records, LLC v. Vimeo, LLC, 972 F. Supp. 2d 537, 549 (S.D.N.Y. 2013) (“[B]ased on the type of music the videos used here—songs by well-known artists, whose names were prominently displayed—and the placement of the songs within the video (played in virtually unaltered form for the entirety of the video), a jury could find that Defendants had ‘red flag’ knowledge of the infringing nature of the videos.”).

[29] Capitol Records, 826 F.3d at 94 (emphasis in original).

[30] Id. at 97.

I presented this at the Fordham IP Conference on April 25, 2019. My oral presentation is here, and the accompanying slides are here.

Categories
Copyright Infringement Internet Uncategorized

Capitol Records v. Vimeo: Courts Should Stop Coddling Bad Actors in Copyright Cases

Here’s a brief excerpt of my new post that was published on IPWatchdog:

Here’s where we are after Capitol Records v. Vimeo: A service provider can encourage its users to infringe on a massive scale, and so long as the infringement it encourages isn’t the specific infringement it gets sued for, it wins on the safe harbor defense at summary judgment. This is so even if there’s copious evidence that its employees viewed and interacted with the specific infringing material at issue. No jury will ever get to weigh all of the evidence and decide whether the infringement is obvious. At the same time, any proactive steps taken by the service provider will potentially open it up to liability for having actual knowledge, so the incentive is to do as little as possible to proactively “detect and deal” with piracy. This is not at all what Congress intended. It lets bad faith service providers trample the rights of copyright owners with impunity.

To read the rest of this post, please visit IPWatchdog.

Categories
Copyright Internet Uncategorized

BMG v. Cox: ISP Liability and the Power of Inference

Cross-posted from the Law Theories blog.

As readers are likely aware, the jury verdict in BMG v. Cox was handed down on December 17th. The jury found that BMG had proved by a preponderance of the evidence that Cox’s users were direct infringers and that Cox is contributorily liable for that infringement. The interesting thing, to me at least, about these findings is that they were both proved by circumstantial evidence. That is, the jury inferred that Cox’s users were direct infringers and that Cox had the requisite knowledge to make it a contributory infringer. Despite all the headlines about smoking-gun emails from Cox’s abuse team, the case really came down a matter of inference.

Direct Infringement of the Public Distribution Right

Section 106(3) grants copyright owners the exclusive right “to distribute copies . . . of the copyrighted work to the public[.]” In the analog days, a copy had to first be made before it could be distributed, and this led to much of the case law focusing on the reproduction right. However, in the digital age, the public distribution usually occurs before the reproduction. In an upload-download scenario, the uploader publicly distributes the work and then the downloader makes the copy. This has brought much more attention to the contours of the public distribution right, and there are some interesting splits in the case law looking at online infringement.

Though from the analog world, there is one case that is potentially binding authority here: Hotaling v. Church of Jesus Christ of Latter-Day Saints. Handed down by the Fourth Circuit in 1997, Hotaling held that “a library distributes a published work . . . when it places an unauthorized copy of the work in its collection, includes the copy in its catalog or index system, and makes the copy available to the public.” The copies at issue in Hotaling were in microfiche form, and they could not be checked out by patrons. This meant that the plaintiff could not prove that the library actually disseminated the work to any member of the public. Guided by equitable concerns, the Fourth Circuit held that “a copyright holder would be prejudiced by a library that does not keep records of public use,” thus allowing the library to “unjustly profit by its own omission.”

Whether this aspect of Hotaling applies in the digital realm has been a point of contention, and the courts have been split on whether a violation of the public distribution right requires actual dissemination. As I’ve written about before, the Nimmer on Copyright treatise now takes the position that “[n]o consummated act of actual distribution need be demonstrated in order to implicate the copyright owner’s distribution right,” but that view has yet to be universally adopted. Regardless, even if actual dissemination is required, Hotaling can be read to stand for the proposition that it can be proved by circumstantial evidence. As one court put it, “Hotaling seems to suggest” that “evidence that a defendant made a copy of a work available to the public might, in conjunction with other circumstantial evidence, support an inference that the copy was likely transferred to a member of the public.”

The arguments made by BMG and Cox hashed out this now-familiar landscape. Cox argued that merely offering a work to the public is not enough: “Section 106(3) makes clear that Congress intended not to include unconsummated transactions.” It then distinguished Hotaling on its facts, suggesting that, unlike the plaintiff there, BMG was “in a position to gather information about alleged infringement, even if [it] chose not to.” In opposition, BMG pointed to district court cases citing Hotaling, as well as to the Nimmer treatise, for the proposition that making available is public distribution simpliciter.

As to Cox’s attempt to distinguish Hotaling on the facts, BMG argued that Cox was the one that failed “to record actual transmissions of infringing works by its subscribers over its network.” Furthermore, BMG argued that “a factfinder can infer that the works at issue were actually shared from the evidence that they were made available,” and it noted that cases Cox had relied on “permit the inference that dissemination actually took place.” In its reply brief, Cox faulted BMG for reading Hotaling so broadly, but it noticeably had nothing to say about the propriety of inferring that dissemination had actually taken place.

In his memorandum opinion issued on December 1st, District Judge Liam O’Grady sided with Cox on the making available issue and with BMG on the permissibility of inference. Reading Hotaling narrowly, Judge O’Grady held that the Fourth Circuit merely “articulated a principle that applies only in cases where it is impossible for a copyright owner to produce proof of actual distribution.” And without the making available theory on the table, “BMG must show an actual dissemination of a copyrighted work.” Nonetheless, Judge O’Grady held that the jury could infer actual dissemination based on the circumstantial evidence collected by BMG’s agent, Rightscorp:

Cox’s argument ignores the fact that BMG may establish direct infringement using circumstantial evidence that gives rise to an inference that Cox account holders or other authorized users accessed its service to directly infringe. . . . Rightscorp claims to have identified 2.5 million instances of Cox users making BMG’s copyrighted works available for download, and Rightscorp itself downloaded approximately 100,000 full copies of BMG’s works using Cox’s service. BMG has presented more than enough evidence to raise a genuine issue of material fact as to whether Cox account holders directly infringed its exclusive rights.

The jury was ultimately swayed by this circumstantial evidence, inferring that BMG had proved that it was more likely than not that Cox’s users had actually disseminated BMG’s copyrighted works. But proving direct infringement is only the first step, and BMG next had to demonstrate that Cox is contributorily liable for that infringement. As we’ll see, this too was proved by inference.

Contributory Infringement of the Public Distribution Right

While the Patent Act explicitly provides circumstances in which someone “shall be liable as a contributory infringer,” the Copyright Act’s approach is much less direct. As I’ve written about before, the entire body of judge-made law concerning secondary liability was imported into the 1976 Act via the phrase “to authorize” in Section 106. Despite missing this flimsy textual hook, the Supreme Court held in Sony that nothing precludes “the imposition of liability for copyright infringements on certain parties who have not themselves engaged in the infringing activity.” Indeed, the Court noted that “the concept of contributory infringement is merely a species of the broader problem of identifying the circumstances in which it is just to hold one individual accountable for the actions of another.”

Arguments about when it’s “just” to hold someone responsible for the infringement committed by another have kept lawyers busy for well over a century. The Second Circuit’s formulation of the contributory liability test in Gershwin has proved particularly influential over the past four decades: “[O]ne who, with knowledge of the infringing activity, induces, causes or materially contributes to the infringing conduct of another, may be held liable as a ‘contributory’ infringer.” This test has two elements: (1) knowledge, and (2) induce, cause, or materially contribute. Of course, going after the service provider, as opposed to going after the individual direct infringers, often makes sense. The Supreme Court noted this truism in Grokster:

When a widely shared service or product is used to commit infringement, it may be impossible to enforce rights in the protected work effectively against all direct infringers, the only practical alternative being to go against the distributor of the copying device for secondary liability on a theory of contributory or vicarious infringement.

And this is what BMG has done here by suing Cox instead of Cox’s users. The Supreme Court in Grokster also introduced a bit of confusion into the contributory infringement analysis. The theory at issue there was inducement—the plaintiffs argued that Grokster induced its users to infringe. Citing Gershwin, the Supreme Court stated this test: “One infringes contributorily by intentionally inducing or encouraging direct infringement[.]” Note how this is narrower than the test in Gershwin, which for the second element also permits causation or material contribution. While, on its face, this can plausibly be read to imply a narrowing of the traditional test for contributory infringement, the better read is that the Court merely mentioned the part of the test (inducement) that it was applying.

Nevertheless, Cox argued here that Grokster jettisoned a century’s worth of the material contribution flavor of contributory infringement: “While some interpret Grokster as creating a distinct inducement theory, the Court was clear: Grokster is the contributory standard.” Cox wanted the narrower inducement test to apply here because BMG would have a much harder time proving inducement over material contribution. As such, Cox focused on its lack of inducing behavior, noting that it did not take “any active steps to foster infringement.”

Despite its insistence that “Grokster supplanted the earlier Gershwin formulation,” Cox nevertheless argued that BMG’s anticipated material contribution claim “fails as a matter of law” since the knowledge element could not be proved. According to Cox, “Rightscorp’s notices do not establish Cox’s actual knowledge of any alleged infringement because notices are merely allegations of infringement[.]” Nor does the fact that it refused to receive notices from Rightscorp make it “willfully blind to copyright infringement on its network.” Cox didn’t argue that its service did not materially contribute to the infringement, and rightfully so—the material contribution element here is a no-brainer.

In opposition, BMG focused on Gershwin, declaring it to be “the controlling test for contributory infringement.” BMG noted that “Cox is unable to cite a single case adopting” its narrow “reading of Grokster, under which it would have silently overruled forty years of contributory infringement case law” applying Gershwin. (Indeed, I have yet to see a single court adopt Cox’s restrictive read of Grokster. This hasn’t stopped defendants from trying, though.) Turning to the material contribution element, BMG pointed out that “Cox does not dispute that it materially contributed to copyright infringement by its subscribers.” Again, Cox didn’t deny material contribution because it couldn’t win on this argument—the dispositive issue here is knowledge.

On the knowledge element, BMG proffered two theories. The first was that Cox is deemed “to have knowledge of infringement on its system where it knows or has reason to know of the infringing activity.” Here, BMG had sent Cox “millions of notices of infringement,” and it argued that Cox could not “avoid knowledge by blacklisting, deleting, or refusing” to accept its notices. Moreover, BMG noted that “Cox’s employees repeatedly acknowledged that they were aware of widespread infringement on Cox’s system.” BMG additionally argued that Cox was willfully blind since it “blacklisted or blocked every single notice of copyright infringement sent by Rightscorp on behalf of Plaintiffs, in an attempt to avoid specific knowledge of any infringement.”

In reply, Cox cited Sony for the rule that “a provider of a technology could not be liable for contributory infringement arising from misuse if the technology is capable of substantial noninfringing uses.” And since Cox’s service “is capable of substantial noninfringing users,” it claimed that it “cannot be liable under Sony.” Of course, as the Supreme Court clarified in Grokster, that is not the proper way to read Sony. Sony merely says that knowledge cannot be imputed because a service has some infringing uses. But BMG here is not asking for knowledge to be imputed based on the design of Cox’s service. It’s asking for knowledge to be inferred from the notices that Cox refused to receive.

Judge O’Grady made short work of Cox’s arguments. He cited Gershwin as the controlling law and rejected Cox’s argument vis-à-vis Grokster: “The Court finds no support for Cox’s reading of Grokster.” In a footnote, he brushed aside any discussion of whether Cox materially contributed to the infringement since Cox failed to raise the point in its initial memorandum. Judge O’Grady then turned to the knowledge element, stating the test as this: “The knowledge requirement is met by a showing of actual or constructive knowledge or by evidence that a defendant took deliberate actions to willfully blind itself to specific infringing activity.” In a footnote, he declined to follow the narrower rule in the Ninth Circuit from Napster that requires the plaintiff to establish “actual knowledge of specific acts of infringement.”

Thus, Judge O’Grady held that three types of knowledge were permissible to establish contributory infringement: (1) actual knowledge (“knew”), (2) constructive knowledge (“had reason to know”), or (3) willful blindness. Rejecting Cox’s theory to the contrary, he held that “DMCA-compliant notices are evidence of knowledge.” The catch here was that Cox refused to receive them, and it even ignored follow-up emails from BMG. And this is where inference came into play: Judge O’Grady held that Cox could have constructive knowledge since “a reasonable jury could conclude that Cox’s refusal to accept Rightscorp’s notices was unreasonable and that additional notice provided to Cox gave it reason to know of the allegedly infringing activity on its network.”

Turning to willful blindness, Judge O’Grady stated that it “requires more than negligence or recklessness.” Citing Global-Tech, he noted that BMG must prove that Cox “took ‘deliberate actions to avoid confirming a high probability of wrongdoing and who can almost be said to have actually known the critical facts.’” The issue here was clouded by the fact that Cox didn’t simply refuse to accept BMG’s notices from Rightscorp, but instead it offered to receive them if certain language offering settlements to Cox’s users was removed. While it would be reasonable to infer that Cox was not “deliberately avoiding knowledge of illegal activity,” Judge O’Grady held that “it is not the only inference available.” As such, he left it for the jury to decide as a question of fact which inference was better.

The jury verdict is now in, and we don’t know whether the jury found for BMG on the constructive knowledge theory or the willful blindness theory—or perhaps even both. Either way, the question boiled down to one of inference, and the jury was able to infer knowledge on Cox’s part. And this brings us back to the power of inference. Cox ended up being found liable as a contributory infringer for its users’ direct infringement of BMG’s public distribution rights, and both of these verdicts were established with nothing more than circumstantial evidence. That’s the power of inference when it comes to ISP liability.