Copyhype

Imagine an author. She has a spark of inspiration and sets out developing a story. Characters are sketched out. Index cards bearing plot points are written up, arranged, and rearranged. As the story takes shape, she calls in her assistant, who, sitting at a computer, types in the first draft of the story as the author dictates.

Now, we ask, who has written the story?

Common sense would tell us the author, of course, has written the story, and few would dispute that answer. But in the strictest, technical sense, if we define “to write” narrowly enough to mean only the last act between thought and word on page, the ministerial act, it is the assistant who wrote the story, by typing it out.

Beyond semantics, that narrow definition is not particularly useful. Only the trivia buff would be interested in the participation of the assistant; most of us want to know who the author was, who is responsible, in the real sense, for the story we are reading. We don’t need a million monkeys to know that the process of writing is more than tapping on a keyboard.

In recent decades, due to technology, copyright law has increasingly had to deal with a similar question. Because copyright law is concerned primarily with the act of copying, some courts have had to ask who has made the copy. When a copy has been made without authorization, who is the author of the infringement?

This question becomes most palpable when we start talking about what I will refer to as “copy machines.” These include literal copiers like photocopiers, tape recorders, or DVRs, as well as virtual copy machines—computing processes, both on local devices or working within the cloud. The problem is that some courts have adhered to the strict technical sense of words and defined “to copy” so narrowly that it doesn’t correspond with the common sense definition.

Aereo

The Supreme Court is set to hear oral arguments in American Broadcasting Companies v. Aereo on April 22. The question of who “caused” the infringement is raised by Aereo in its brief, though the issue was not addressed by either the district or appellate courts below. However, it does cast a long shadow over the proceedings. Some commenters have raised the fear that the entire fate of the cloud computing industry rests on the Supreme Court’s decision (this year’s iteration of “copyright will break the internet,” I suppose). ¹See, for example, David Sohn, Cloud Computing Threatened in Aereo Supreme Court Case, CDT (March 3, 2014); Marc Perton, Cablevision: Case against Aereo could destroy cloud computing (and our cloud DVR), Engadget (Dec. 13, 2013).

This argument is unsound for at least two reasons. First, the Cablevision decision, upon which Aereo relies on, only applies within the Second Circuit, so the fact that cloud computing has flourished in other parts of the US, and throughout the world, suggests the decision is not the panacea Aereo supporters claim it is. ²Cloud computing does not, for example, appear to have been destroyed in Australia despite a 2012 decision that is the equivalent of reversing Aereo. Second, service providers already have safe harbor from liability for infringing activity stemming from user-directed storage. ³17 USC § 512(c). So any protection Supreme Court affirmation would provide would largely be redundant.

Nevertheless, US case law has yet to fully develop a satisfactory answer to the question of who “makes” a copy, and it is safe to say the issue will maintain its prominence in the near future. In this article, I want to take a look at this case law to see how courts have approached the question, which is not a novel question in the law. Particularly, I look at the emergence of the “volitional conduct” test to mediate disputes over direct liability on the internet. I note that most often, this test resembles the legal doctrine of proximate causation and argue that ditching volitional conduct entirely in favor of a more direct focus on proximate causation would offer more clarity and better results. I finish by sketching out what a proximate causation inquiry might look like.

Copying, causation, and volition in the courts

Prior to the popularization of the internet, a handful of cases examined the contours of liability for those who make, own, or operate copy machines utilized by others.

Make-A-Tape

In 1973, a court considered a chain of electronics stores operating “Make-A-Tape” machines. ⁴Elektra Records v. Gem Electronic Distributors, 360 F. Supp. 821 (EDNY 1973). Customers could select an album from the store’s catalog, purchase one of the blank tapes sold by the store, conveniently receiving as change the two quarters required to operate the machine, and start the “Make-A-Tape.” Within two minutes, the customer could leave the store with a perfect copy of the album at a fraction of the cost.

The Eastern District Court of New York did not delve too deeply into the question of causation, but it did conclude that “Regardless of the precise role played by defendants’ employees, the above-described operation of the Make-A-Tapes clearly evidences their commercial exploitation by defendants for profit in derogation of plaintiffs’ rights of exclusive publication.”

The defendants attempted to escape direct liability by focusing the court’s attention on the self-service nature of the copiers and their resemblance to “a photocopier in a public library.” This did not convince the court, which focused on the differences in time, quality, and cost of reproducing entire works on a photocopier compared to the Make-A-Tape machine. It also pointed out the non-profit nature and altruistic motives of a library, which differ from how the defendants “manifestly utilize the Make-A-Tape as a further source of income.”

Sony Betamax

The issue of who makes a copy did not, apparently, come in front of courts again for another six years, in Universal City Studios v. Sony Corp. of America. The Supreme Court’s decision regarding the secondary liability of a Betamax manufacturer is considered a cornerstone of modern copyright jurisprudence. But little attention is paid to the fact that, years before the highest court would consider indirect liability, the district court considered whether Sony should be directly liable for infringement of Betamax users.

The district court noted that “It is true that one can be found to have infringed directly even without participating in the actual infringing activity.” But the court ultimately held Sony could not be directly liable for copying occurring on the machines it designs and distributes for a number of reasons.

First, Sony does not “loan or otherwise provide the copyrighted work” that is being copied. Second, the copying does not occur on premises “operated and managed by the defendants but rather in a person’s home, a location in which individual privacy is constitutionally protected and over which defendants have no control.” ⁵The Ninth Circuit would later hold that a place “open to the public” does not include a hotel room because an individual renting a hotel room enjoys a similar constitutional right of privacy. Columbia Pictures Industries, Inc. v. Professional Real Estate Inv., Inc., 866 F. 2d 278, 281 (9th Cir. 1989). Finally, the Betamax players aren’t solely used for unauthorized copying; they have a range of uses and can be used for noninfringing purposes. ⁶The Supreme Court would eventually hold that this characteristic of Betamax players also prevents the imputation of requisite knowledge for contributory infringement by Sony.

Kinko’s

Print shop Kinko’s would find itself in court in the early 90’s, the subject of an infringement complaint for operating a “course packet” service—college professors would provide nearby Kinko’s with selections and chapters from original copyrighted works that had been assigned to students, and Kinko’s would duplicate and assemble the materials into anthologies that they would then sell to students. ⁷Basic Books, Inc. v. Kinko’s Graphics Corp., 758 F. Supp. 1522 (SDNY 1991).

There was little argument that Kinko’s was “making” the copies. However, when it came to assessing damages, Kinko’s argued that it was acting as an agent of the professors’ college since “Section 504(c) provides that the court ‘shall remit statutory damages … where an infringer believed and had reasonable grounds for believing that his or her use of the copyrighted work was a fair use …, if the infringer was (i) an employee or agent of a nonprofit educational institution….’ 17 U.S.C. § 504(c)(2).”

The court rejected this argument, saying there was no evidence of an agency relationship between Kinko’s and the university, and Kinko’s was unable to show that “the professors exerted a sufficient level of control over the relationship.”

Frena

Finally, one of the earliest cases involving copyright liability in the online context saw a bulletin board system (BBS) operator liable for directly infringing public display and distribution rights. ⁸Playboy Enterprises v. Frena, 839 F. Supp. 1552 (M.D. Fla. 1993). Though there was only brief discussion in Frena, the court did say, “There is no dispute that Defendant Frena supplied a product containing unauthorized copies of a copyrighted work. It does not matter that Defendant Frena claims he did not make the copies itself.”

Netcom

And then came Netcom. In 1995, the Northern District Court of California was confronted with the question of copyright liability for an internet access provider based on infringement occurring on Usenet groups. ⁹Religious Tech. Center v. Netcom On-line Comm., 907 F. Supp. 1361 (ND Cali. 1995). Usenet is a decentralized communications system consisting of thousands of “newsgroups” where users can read and post public messages. Internet access providers can set up a Usenet server to become part of the system and allow their subscribers to interact with Usenet, but the system otherwise functions independently from any one internet access provider—indeed, it operates independently from any individual or entity.

A critic of the Church of Scientology had posted a number of unpublished and published works, without permission, of Scientology founder L. Ron Hubbard to a religion newsgroup. The copyright holders of the works sued for copyright infringement, naming the critic who uploaded the works, the operator of the BBS where the works were uploaded, and Netcom, which provided internet access to the BBS (and, by extension, to the critic). Netcom moved for summary judgment on the issue of its own direct liability.

The court granted the motion, saying “Although copyright is a strict liability statute, there should still be some element of volition or causation which is lacking where a defendant’s system is merely used to create a copy by a third party.” The court described the copying done by Netcom as “automatic[] and uniform[]” and its actions as “necessary to have a working system for transmitting Usenet postings to and from the Internet.” Thus, “the mere fact that Netcom’s system incidentally makes temporary copies of plaintiffs’ works does not mean Netcom has caused the copying.”

Netcom would serve as the stepping off point for analyzing direct liability online. Nearly every subsequent case where the issue of direct liability for the owner or operator of a copying service or platform arose refers to the case.

But while these courts agree that Netcom said direct liability requires some form of “volitional conduct”, there is little agreement over what that means.

After Netcom

The Fourth Circuit adopted Netcom in CoStar Group v. LoopNet, saying

to establish direct liability under §§ 501 and 106 of the Act, something more must be shown than mere ownership of a machine used by others to make illegal copies. There must be actual infringing conduct with a nexus sufficiently close and causal to the illegal copying that one could conclude that the machine owner himself trespassed on the exclusive domain of the copyright owner. The Netcom court described this nexus as requiring some aspect of volition or causation. ¹⁰373 F.3d 544, 550 (4th Cir. 2004).

However, the language of CoStar, rather than describing a general causation principle, suggests that it was laying down a blackletter rule that “an ISP who owns an electronic facility that responds automatically to users’ input is not a direct infringer.” A subsequent, unpublished Fourth Circuit opinion lends credence to this notion; in Quantum Systems Integrators v. Sprint Nextel, the court quoted CoStar‘s holding “that ISPs, when passively storing material at the direction of users in order to make that material available to other users upon their request, do not ‘copy’ the material in direct violation of § 106 of the Copyright Act” and rejected defendant’s argument that it was shielded from liability under this rule in part by pointing out that it was not engaging “in ‘conduct typically engaged in by an ISP.'” ¹¹338 Fed. Appx. 329 (4th Cir. 2009).

In Cartoon Network v. CSC Holdings (“Cablevision“), the Second Circuit adopted Netcom, holding that copies produced by a cable provider’s remote DVR (RS-DVR) system are “made” by the cable provider’s customers, and the cable provider’s contribution of providing the system doesn’t warrant the imposition of direct liability. ¹²536 F. 3d 121 (2nd Circuit 2008). Unlike CoStar, Cablevision didn’t read Netcom as establishing a rule for analyzing the liability of ISPs—indeed, Cablevision operated a sui generis service wholly separate from the Internet. Instead, Cablevision read into Netcom a distinction between human acts and technical processes, with only the former supplying the necessary “volition” for direct liability. “In determining who actually ‘makes’ a copy,” said the Second Circuit, “a significant difference exists between making a request to a human employee, who then volitionally operates the copying system to make the copy, and issuing a command directly to a system, which automatically obeys commands and engages in no volitional conduct.”

The Cablevision court never explains exactly what this difference is or why it is so significant. (Indeed, this statement seems contrary to the Fourth Circuit’s claim in CoStar that employee review of images before they are posted “tends only to lessen the possibility that LoopNet’s automatic electronic responses will inadvertently enable others to trespass on a copyright owner’s rights.” As noted later on, some courts have been too quick to read causation out of the design of automated processes.)

Other cases have taken the same view as Cablevision and held that Netcom stands for the proposition that the design of automated technical processes is never causational. Some courts describe this rule as saying the lack of human intervention means there is no volitional conduct. ¹³Perfect 10 v. Giganews, CV11-07098 AHM (SHx) (CD Cali, March 8, 2013); Wolk v. Kodak Imaging Network, Inc., 840 F. Supp.2d 724, 742 (SDNY 2012), “There is no dispute that any reproduction, display or transmission of the Plaintiff’s images by or through the KODAK Gallery website is an automated process with no human intervention by any employee of the Kodak Defendants. The fact that Wolk’s images are copied into product simulations in addition to being transmitted to fulfillment vendors does not constitute a volitional act where the copying is automated”; Disney Enterprises v. Hotfile, 798 F. Supp. 2d 1303, 1309 (SD Fla 2011), “courts have repeatedly held that the automatic conduct of software, unaided by human intervention, is not ‘volitional'”; Parker v. Google, Inc., 422 F. Supp. 2d 492 (ED Pa 2006), “When an ISP automatically and temporarily stores data without human intervention so that the system can operate and transmit data to its users, the necessary element of volition is missing. The automatic activity of Google’s search engine is analogous”; Field v. Google Inc., 412 F. Supp. 2d 1106 (D. Nevada 2006), holding that, regarding the creation and downloading of a cached copy of a web page from Google’s servers, “The automated, non-volitional conduct by Google in response to a user’s request does not constitute direct infringement under the Copyright Act.” But that is not entirely true. Humans certainly intervened, volitionally, to design the system and maintain and operate it. So, contrary to how courts here have described it, it is more accurate to say this permutation of the “volitional conduct” test limits direct liability when there is no human intervention contemporaneous with the moment a copy is initiated.

Finally, some courts have interpreted “volitional conduct” as roughly synonymous with the general tort principle of “proximate causation.”

At least one court has explicitly called attention to this resemblance. “Notwithstanding these sound decisions, the concept of ‘volition’ can be confusing,” said the Central District Court of California. “‘Volitional’ is sometimes understood to mean ‘intentional,’ and yet no showing of intent is required for direct infringement liability. In this Court’s view, the key to understanding the so-called ‘volitional conduct’ requirement is to equate it with the requirement of causation, not intent. ‘Just who caused the copyrighted material to be infringed?'” ¹⁴Perfect 10 v. Giganews, supra.

And the Ninth Circuit seems to have implicitly taken the position that volitional conduct is equivalent to proximate causation.

In Fox Broadcasting Co. v. Dish Networks, satellite TV provider Dish was sued, in part, for unauthorized copying of television programs made through its PrimeTime Anytime Service (PTAT). When enabled by a subscriber, PTAT automatically recorded all primetime programs from the four major broadcasters and made them available on the subscriber ‘s DVR. Dish argued that its subscribers should be the ones exposed to liability because they, not Dish, “made” the copies.

The Central District Court of California agreed with Dish. On a preliminary injunction motion, the court said Dish was likely to prevail on the merits, relying primarily on the holdings of Netcom and Cablevision that require some element of “volitional conduct” for direct liability.

The decision was appealed, and the Ninth Circuit affirmed. But there’s something curious about the opinion. The Circuit does not once mention volition, nor does it cite to the granddaddy of volitional conduct, Netcom (though it does cite approvingly to Cablevision). Instead, it said simply that direct liability “comprises a requirement that the defendant cause the copying.” (Emphasis added). In addition, the court repeats the lower court’s quotation to Prosser’s recitation of the proximate causation inquiry. ¹⁵The lower court initially misquoted Prosser as saying the proximate causation inquiry looked at who was “the most significant and important cause” of the copy rather than “whether the conduct has been so significant and important a cause that the defendant should be legally responsible” a fact seized upon by plaintiff’s in their motion for rehearing en banc. The Ninth Circuit denied the rehearing but amended the original opinion to note the misquotation, though the ultimate outcome remained unchanged. At least one district court within the Ninth Circuit has since suggested that the Circuit had yet to adopt the “volitional conduct” test. ¹⁶National Photo Group v. Allvoices, No. C-13-03627 JSC, n.3 (ND Cali, Jan. 24, 2014).

In short, it is not remarkable to say “that the principle of volition has been applied inconsistently and is not always well understood.” ¹⁷Eleanor M. Lackman and Scott J. Sholder, The Role of Volition in Evaluating Direct Copyright Infringement Claims Against Technology Providers, 22 Bright Ideas 3 (2013). The above discussion demonstrates that, since Netcom, at least three versions of the “volitional conduct” test have emerged (four, if you count cases where its application was declined entirely).

I think the last view—that volitional conduct is merely a substitute for proximate causation—is correct, and that the different terminology has led courts astray. Focusing on proximate causation instead of volitional conduct would provide better results. To see why, let’s take a closer look at proximate causation.

Proximate Causation and Copyright

Copyright infringement is a strict liability tort. Knowledge or intent are not required for copyright infringement. ¹⁸Sater Design Collection v. Waccamaw Construction, No. 4:08-CV-4133-TLW-SVH (D. S.C., Feb. 14, 2011) (“Case law establishes that knowledge or intent is not an element of copyright infringement “). But infringement does require, like all torts, some element of causation, an act that results in the harm. Causation is typically divided into factual (“but for”) causation and “proximate” causation. Put another way, did the act actually cause the harm and, if so, should the law, as a matter of policy, hold the actor liable for causing the harm. ¹⁹“Causation principles generally applicable to tort liability must be considered applicable. These require not only cause-in-fact, but “legal” or “proximate” cause as well, the latter involving a policy rather than a purely factual determination: “whether the conduct has been so significant and important a cause that the defendant should be held responsible.” Brandenburg v. Seidel, 859 F. 2d 1179, 1189 (4th Cir. 1988), quoting Prosser and Keeton Torts, § 42 p. 272 (general principle) (5th ed. 1984); “Cause in fact, “but for” causation, is not enough for liability. ” * * * `”Once it is established that the defendant’s conduct has in fact been a cause of the injury * * * there remains the question whether the defendant should be legally responsible for what he has caused * * *,”` ” and that is a question of policy to be resolved by the courts.” Benner v. Bell, 602 NE 2d 896, 899 (Ill App 4d 1992).

Factual causation is not typically difficult to establish since it is such a broad concept. Proximate causation, on the other hand, is an entirely different story. It is not a philosophical or scientific question, but a legal one.

“What we do mean by the word “proximate” is that, because of convenience, of public policy, of a rough sense of justice, the law arbitrarily declines to trace a series of events beyond a certain point.” ²⁰Palsgraf v. Long Island Railroad Co., 248 N.Y. 339 (1928) (Andrews, J., dissenting).

The factual causes of any occurrence are seemingly limitless. One could trace forever the chain of events that lead to a specific harm, but it has long been recognized that there is little sense in extending legal responsibility throughout those chains. The chief mechanism animating this recognition is proximate causation. Proximate causation is a limiting theory about the “specific qualities that an agency must possess in relation to the outcome in order to be its cause in law.” ²¹Antony Honoré, “Causation in the Law”, The Stanford Encyclopedia of Philosophy (Winter 2010 Edition), Edward N. Zalta (ed.). As Chief Justice Roberts has said, proximate cause “limits liability at some point before the want of a nail leads to loss of the kingdom.” ²²CSX Transportation v. McBride, 131 S. Ct. 2630, 2646 (2011).

Justice O’Connor, in a case concerning causation under an environmental statute that is silent on the issue, explains how proximate causation is an element of all strict liability torts and roughly sketches its contours and purpose:

Strict liability means liability without regard to fault; it does not normally mean liability for every consequence, however remote, of one’s conduct. I would not lightly assume that Congress, in enacting a strict liability statute that is silent on the causation question, has dispensed with this well-entrenched principle. In the absence of congressional abrogation of traditional principles of causation, then, private parties should be held liable under § 1540(a)(1) only if their habitat-modifying actions proximately cause death or injury to protected animals…

Proximate causation is not a concept susceptible of precise definition. It is easy enough, of course, to identify the extremes. The farmer whose fertilizer is lifted by a tornado from tilled fields and deposited miles away in a wildlife refuge cannot, by any stretch of the term, be considered the proximate cause of death or injury to protected species occasioned thereby. At the same time, the landowner who drains a pond on his property, killing endangered fish in the process, would likely satisfy any formulation of the principle. We have recently said that proximate causation “normally eliminates the bizarre,” and have noted its “functionally equivalent” alternative characterizations in terms of foreseeability, and duty, Proximate causation depends to a great extent on considerations of the fairness of imposing liability for remote consequences. The task of determining whether proximate causation exists in the limitless fact patterns sure to arise is best left to lower courts. ²³Babbitt v. Sweet Home Chapter, Communities for Great Ore., 515 US 687, 712-13 (1995) (O’Connor concurrence).

O’Connor notes the role of fairness in the proximate causation inquiry; other commenters point out the role of justice, blame, and moral responsibility. Some common policy questions underlying proximate causation include “indications of preconceived purpose, specifically intended consequence, necessary or natural result, reasonable foreseeability of result, the intervention of independent causes, whether the defendant’s acts are a substantial factor in the sequence of responsible causation, and the factual directness of the causal connection.” ²⁴Khurana v. Innovative Health Care Systems, Inc., 130 F. 3d 143, 148-49 (5th Cir. 1997). See also Benner v. Bell, supra at 899, noting policy questions underlying proximate causation include “reasonable foreseeability of injury, likelihood of injury, magnitude of burden of guarding against it, and consequences of placing that burden on defendant.”

Proximate causation is often tied to our expectations. “One connotation of proximate cause is that harm came about in a ‘direct’ or expected way, rather than in an unusual, freakish manner.” In addition, liability may not be found “because of lack of proximate cause in the sense that the accidents are coincidental to defendants’ behavior, unrelated to the normal risk created by their behavior.” ²⁵Steven Shavell, “Causation and Tort Liability “.

Finally, proximate causation contains a recognition that liability must necessarily cut off. “At some point in the causal chain, the defendant’s conduct or product may be too remotely connected with the plaintiff’s injury to constitute legal causation ” ²⁶Union Pump Co. v. Allbritton, 898 SW 2d 773, 775-76 (Texas 1995).

If we revisit the Netcom decision through the lens of proximate causation, we can easily see that the court was considering the same policies. For instance, it considers the consequences of placing liability on the defendant when it notes that holding Netcom liable “would create many separate acts of infringement and, carried to its natural extreme, would lead to unreasonable liability.” It looks at whether the defendant’s acts are a substantial factor in the sequence of responsible causation and the magnitude of the burden of guarding against it when it says holding Netcom directly liable “would hold the entire Internet liable for activities that cannot reasonably be deterred.” The court is, in other words, demonstrating reasons it should limit Netcom’s legal liability despite its factual role in causing the copying.

That is to say that when Netcom says direct liability requires “some element of volition or causation”, it is referring primarily to the latter, not the former. Read appropriately, Netcom means “proximate causation” when it says “volitional conduct”.

If we agree that the “volitional conduct” test and proximate causation are largely synonymous than we can, and should, ditch any talk of “volitional conduct.”

This is preferable for a number of reasons. For one thing, the “volitional conduct” test provides no advantage over a proximate causation inquiry.

Saying direct liability requires volitional conduct is redundant. The second Restatement of the Law on Torts says that an intentional tort requires an “act”. It defines an “act” as “an external manifestation of the actor’s will” and notes “there cannot be an act without volition”. So to say that copyright infringement requires a volitional act is simply to repeat the obvious and doesn’t get us far. It is like saying we have a right to “free speech on Tuesdays.” ²⁷See Frederick Schauer, Free Speech on Tuesdays, Virginia Public Law and Legal Theory Research Paper No. 2014-10 (January 10, 2014). Schauer takes on the question of what it means to say a right “exists” and examines the way we talk about—and specify—rights. Noting that “there is no remark without remarkability,” Schauer writes that, “although it is thus true that there is in fact a right to free speech on Tuesdays, it would be odd for someone to say that there is a right to free speech on Tuesdays. And that is because to say that there is a right to free speech on Tuesdays implies that there is something different or special about free speech on Tuesdays as compared to some background or baseline understand.” Similarly, to say that causation requires volitional conduct is both technically true but implicitly “remarkable.”

And, in fact, it puts courts and litigants at a disadvantage. Calling proximate causation “volitional conduct” adds little or no clarity or guidance to the test and only further masks the policies that drive proximate causation analysis in the first place.

Proximate causation itself is already a proxy for an unspecified set of policies that underlie law. By skipping over analysis of proximate causation, courts miss the opportunity to be informed by those policies. At the same time, there have been little or no underlying policies that have emerged as guidance since Netcom for applying the volitional conduct test.

At its core, the question of who “makes” a copy when we are dealing with owners, operators or other intermediaries of copy machines is this: at what point do we circumscribe legal responsibility for parties who are factual causes of a copy. That is precisely the type of question that proximate causation seeks to answer.

One initial result of this decision is that courts need to take care in per se rules regarding automated processes. In a world with self-driving cars and computers that can play chess, it should be apparent that the mere fact that an act occurs via an automated function that can operate independent of the operator should not preclude the operator or programmer of the function as a causal agent of the consequences. Whether an automated process is a causal agent of the operator is a question of fact, but surely the design or programming of a process is, in the language of tort law, an external manifestation of the programmer’s will.

Indeed, you rarely see cases outside the copyright context reflect this sort of thinking. For example, just last month, a federal court held that search engine results are protected by the First Amendment. During its analysis, the court stated that “the fact that search-engine results may be produced algorithmically” is irrelevant. “After all, the algorithms themselves were written by human beings, and they ‘inherently incorporate the search engine company engineers’ judgments about what material users are most likely to find responsive to their queries.'” It would be weird to say the results of an automated process can be considered protected speech attributed to the operator of the system employing that process, unless they are infringing, in which case they are not attributable to the operator. And indeed, at least a few courts have bucked the trend and found the design of automated processes indicative of causation. ²⁸See, for example, Capitol Records, LLC v. ReDigi, Inc., 934 F. Supp. 2d 640, 657 (SDNY 2013), “While that process is itself automated, absolving ReDigi of direct liability on that ground alone would be a distinction without a difference. The fact that ReDigi’s founders programmed their software to choose copyrighted content satisfies the volitional conduct requirement and renders ReDigi’s case indistinguishable from those where human review of content gave rise to direct liability.”

Analyzing Proximate Causation

The ultimate conclusion here is that using proximate causation to answer the question of who makes a copy is sound. It is superior to the ill-defined “volitional conduct” test that some courts have used and should provide better guidance for courts to reach preferable outcomes. So what would a proximate causation analysis look like?

A full dissertation on proximate causation of copying online is beyond this article, but I think it is still worth attempting to trace out some rudimentary factors that courts might consider.

Since we are concerned with who makes a copy, it makes sense to focus much of our attention on the various steps and elements needed to create a copy. For example:

1) Location of equipment

Location is important, just as it is in most areas of the law. This is not so much a question of geography, but rather over the nature of the place, particularly who has dominion over the premises. Saying, as Aereo has, that the only difference between it and a personal antenna is the “length of the cable” is as inaccurate and meaningless as saying the only difference between my apartment and the Italian restaurant down the street is the amount of steps I have to take to get there. There are, it turns out, quite a few other differences, most far more relevant to any legal issue that may arise. I would point out that I am not as convinced as some courts are of the analogy between a self-service copy machine and that same machine in someone’s home. The copy machine in someone’s home is reserved for their use and so the amount of copying is necessarily limited. But the copy machine at a business is open to a potentially continuous stream of people coming to copy. ²⁹Though the court in Columbia Pictures Indus. v. Redd Horne, Inc., 568 F. Supp. 494 (WD Pennsylvania 1983), aff’d 749 F. 2d 154 (3rd Cir. 1984), was interpreting the definition of a “place open to the public”, it spoke to this distinction as well. In holding that performances in private viewing booths in a public video store were within the scope of a copyright owner’s exclusive rights, the court noted that “the potential exists for a substantial portion of the public to attend such performances over a period of time.” Id. at 500.

How has this factor been handled by courts? On the one hand, there is little weight given to the fact that copying equipment is on a defendant’s premises—in fact, CoStar expressly exempts the housing and automated operation of copying equipment without more from direct liability. On the other hand, when the copying equipment is outside a defendant’s dominion, courts tend to view defendant’s participation as too remote for direct liability. ³⁰See, for example, In Re Cellco, 663 F. Supp. 2d 363, 377 (SDNY 2009), “Once the customer has downloaded the ringtone onto her telephone, she controls the telephone and makes the decisions that determine whether that ringtone will be triggered by an incoming call signal”; Perfect 10 v Cybernet, 213 F. Supp. 2d 1146, 1168-69 (CD Cali 2002), “Based on the evidence before the Court it appears that Cybernet does not use its hardware to either store the infringing images or move them from one location to another for display. This technical separation between its facilities and those of its webmasters prevents Cybernet from engaging in reproduction or distribution, and makes it doubtful that Cybernet publicly displays the works”; Universal City Studios v. Sony Corp, 480 F. Supp. 429 (CD Cali 1979), “The copying occurs not in a store operated and managed by the defendants but rather in a person’s home, a location in which individual privacy is constitutionally protected and over which defendants have no control.”

2) Provision of copyrighted work

Implicit in many cases relying on Netcom is the fact that the services play a “hands-off” role in the copying, acting as a conduit or providing instead a platform for users to make their own copies—from their own original works. ³¹See, for example, Disney v. Hotfile, 798 F. Supp. 2d 1303, 1307 (SD Fla. 2011), “[N]othing in the complaint alleges that Hotfile … took direct, volitional steps to violate the plaintiffs’ infringement. There are no allegations, say, that Hotfile uploaded copyrighted material.” It makes less sense to limit liability for causing a copy when a service is providing the user with the work to be copied.

Even Cablevision, which found no direct liability for the provider of a remote DVR service, grudgingly admitted this point. The Second Circuit there said Cablevision’s “unfettered discretion in selecting the programming that it would make available for recording… is indeed more proximate to the creation of illegal copying than, say, operating an ISP or opening a copy shop, where all copied content was supplied by the customers themselves or other third parties.”

Courts should be skeptical of attempts by services to argue they are formally not the provider of a work despite functionally providing those works. One example of this can be seen in Columbia Pictures Industries v. Redd Horne. ³²749 F.2d 154 (3rd Cir). Defendant there, sued for infringing plaintiff’s public performance rights for operating viewing booths in its video rental store premises, argued that it was protected by the first sale doctrine. That is, since it had rented the video tapes to its customers, it was no longer responsible for any performances that resulted. The court rejected this argument outright, saying that the first sale doctrine only circumscribes a copyright owner’s control over further distributions. It does not operate as a waiver to a copyright owner’s other exclusive rights such as the public performance right.

But the court also noted that defendant’s argument elevated form over function. “The record clearly demonstrates that showcasing a video cassette at Maxwell’s is a significantly different transaction than leasing a tape for home use. Maxwell’s never disposed of the tapes in its showcasing operations, nor did the tapes ever leave the store. At all times, Maxwell’s maintained physical dominion and control over the tapes.”

Citing to Redd Horne, the court in Warner Bros. WTV Systems rejected a similar argument. There, defendant had built a convoluted internet streaming service where films were kept on DVDs and played on DVD players before being transmitted to user’s computers, rather than storing the digital files on a server. It argued that this set-up made it a service that offered “DVD rentals” rather than public performances, thus taking it outside the scope of copyright law. But, as in Redd Horne, the court found the DVDs remained under physical dominion and control of defendants, and did not resemble a DVD rental service.

3) Participation in copying process

How materially did the defendant participate in the copying process? This is one of those questions that can either threaten to swallow every other consideration or become redundant and act as a sort of meta-factor in the analysis. But I think case law does offer some guidance that can help shape this element into something meaningful and useful.

One example can be found in distinctions between cases involving Usenet services—interestingly, Netcom and a surprisingly large percentage of volitional conduct cases have dealt with Usenet. And, like Netcom, the majority have found that causation “is lacking where a defendant’s system is merely used to create a copy by a third party.” ³³Perfect 10 v. Giganews, CV11-07098 AHM (SHx) (CD Cali, March 8, 2013); Parker v. Google, 422 F. Supp. 2d 492 (ED Pa. 2006); Ellison v. Robertson, 357 F. 3d 1072 (9th Cir. 2004); ALS Scan v. Remarq Communities, 239 F. 3d 619 (4th Cir. 2001). But at least one Usenet service had been found directly liable by a court, and the differences with that service are illustrative here.

In Arista Records v. Usenet.com, the court found, unlike other Usenet services, the defendant engaged in direct infringement by doing the following: it took active measures to create servers dedicated to mp3 files and increased retention time of those files knowing they were most popular content on service, it took active steps including automated filters and human review to remove access to certain categories of content and block certain users, and it routinely exercised control over which newsgroups to accept and store and which to reject.

Each of these steps make Usenet.com more directly a cause of the resulting unauthorized copying. Usenet.com’s actions make infringement a more likely and natural result, thus weakening the claim to limit Usenet.com, a “but for” cause of infringement, as a legal cause of infringement.

A second example of a court examining the participation of a defendant in the copying process is found in Capitol Records v. ReDigi, involving a service purportedly allowing “used” digital music files to be sold. Here the court found it relevant to direct liability that the defendant built a service where only copyrighted work could be sold, despite being automated. It also played a “fundamental and deliberate role” in distribution, including providing infrastructure for users’ infringing sales and affirmatively brokering such sales by connecting interested users with available sellers.

Third, one court said a filelocker could face direct liability for the following reasons: it designed and maintained a system that allows uploading and downloading; it created distinct websites, presumably in an effort to streamline users’ access to different types of media; it encouraged and, in some cases paid, its users to upload vast amounts of popular media through a reward program; it disseminated URLs for various files throughout the internet; it provided payouts to affiliate websites who maintain a catalogue of all available files; and, at a minimum, it was plausibly aware of the ongoing, rampant infringement taking place on its websites. ³⁴Perfect 10 v. Megaupload, 2011 WL 3203117 (S.D. Cal.).

Finally, by way of contrast, the district court in Sony explained a situation where a defendant’s participation was too attenuated to support direct liability. As noted above, the manufacturer of the Betamax maintained no role in copying once the device was sold to a customer and taken to his home. The sale severed all ties with the customer, the device, and any resulting copies.

What can be gleaned from these cases is that evidence that a service provider has not taken a “hands off” or passive role in operating a copying system will defeat limitation on direct liability. That evidence may take the form of intent, purpose, expectation, or ex ante knowledge.

4) Range of uses of “copy machine”

Closely related to the participation in the copying is the range of uses of a copy machine. A defendant who designs a system that is good for nothing but making infringing copies is very proximate to the resulting infringement. The operator of a service with substantial non-infringing uses—a true general-purpose system—is less so. The range necessary narrows as a defendant’s participation in process attenuates—so only devices which are good for pretty much nothing but infringement would give rise to liability as articles of commerce to the manufacturer. ³⁵See, e.g., Sony, supra. Essentially, this factor asks how probable the occurrence of infringement is due to the design of a defendant’s system.

Some courts and commenters single out the “unusual result” as demonstrative of causation, the thing that “makes the difference”. As one scholar explains, “The train wreck is said to be caused by the bent rail even though it was the bent rail together with the fact that the train was going at a certain speed (and other facts, of course) that led to the wreck. To cite the speed—when it is the normal speed of the train—as the cause would be wrong precisely because it is normal and therefore present as a factor when trains are not wrecked.” ³⁶Morton White, Causation in the Law by H. L. A. Hart; A. M. Honoré Review, 60 Columbia Law Review 1058, 1059 (1960). Infringement that occurs on a system designed more for infringing uses is a less unusual result than infringement that occurs on a more general purpose system.

5) Dominion over copy

The party that retains dominion over or possession of a copy once it has been made has less of a claim that he should not be legally responsible for making the copy in the first place. Courts are hesitant to place much stock in the fact that a particular defendant may retain possession over a copy—in Netcom, for example, the court rejected the argument that infringing copies remained on Netcom servers for up to eleven days prohibited it from limiting direct liability. But the converse—a service does not possess the copy once it’s made—has been a factor in limiting direct liability for some courts. ³⁷See In re Cellco, supra.; Cybernet, supra.; Sony, supra.

Dominion over a copy, that is, the right or ability to control the copy, is also relevant. Several courts have limited direct liability by relying on the lack of an ability, legal or technical, to control the copy once it has been made. ³⁸E.g., Leonard v. Stemtech Health Sciences, Inc., CV 08-67-LPS-CJB, 2013 WL 5288266 (D. Del. Sept. 19, 2013), though infringing photo was copied onto a website defendant owned, hosted, and directly profited from, the photo was copied by independent distributor onto an area of the website the defendant did not control. In Dish, though the Ninth Circuit ultimately held that the district court did not abuse its discretion when it denied a preliminary injunction, it did suggest that Dish’s dominion over the copies being made might end up supporting direct liability. The court said, “That Dish decides how long copies are available for viewing, modifies the start and end times of the primetime block, and prevents a user from stopping a recording might be relevant to a secondary or perhaps even a direct infringement claim.”

Miscellany

“Drawing the line between where legal causation may exist and where, as a matter of law, it cannot, has generated a considerable body of law.” ³⁹Union Pump Co. v. Allbritton, 898 SW 2d 773, 775-76 (Texas 1995). Proximate causation is a factual inquiry—a question for the jury. Even then, it is an amorphous concept. The Seventh Circuit, for example, does not even have a definition for proximate causation in its pattern civil jury instructions. ⁴⁰Federal Civil Jury Instructions of the Seventh Circuit (2009). This article argues that courts should resist further use of the “volitional conduct” test to determine liability for services that make copies for users and allow users to make copies and instead rely on the existing tort doctrine of proximate causation, as this would be a clearer and more common sense approach. It does not purport to give any definitive answers to the proximate causation inquiry, nor suggest that doing so would be any easier than applying proximate causation outside the copyright context.

But I do want to end with a couple miscellaneous points worth mentioning.

First, and perhaps most importantly, while there may be similar principles governing causation of each of the various exclusive rights of a copyright owner, courts must be careful about the distinctions between each specific right. Throughout this article, I have used the term “copying” as shorthand for any of the exclusive rights of a copyright owner. ⁴¹SOS, Inc. v. Payday, Inc., 886 F. 2d 1081, 1085 n.3 (9th Cir. 1989); Playboy Enterprises, Inc. v. Webbworld, Inc., 991 F. Supp. 543, 550-51 (ND Tx 1997), “‘Copying’ is a judicial shorthand for the infringement of any of a copyright owner’s exclusive rights.” However, the factors I trace are most relevant to the reproduction right. The elements of an act of reproduction are different from the elements of an act of distribution or a public performance. To note one important difference: when a copy of a work is made, there is only a single act of reproduction. A transmission, by contrast, may involve multiple performances. ⁴²See House Report No. 94-1476 (1976), “Under the definitions of “perform,” “display,” “publicly,” and “transmit” in section 101, the concepts of public performance and public display cover not only the initial rendition or showing, but also any further act by which that rendition or showing is transmitted or communicated to the public. Thus, for example: a singer is performing when he or she sings a song; a broadcasting network is performing when it transmits his or her performance (whether simultaneously or from records); a local broadcaster is performing when it transmits the network broadcast; a cable television system is performing when it retransmits the broadcast to its subscribers; and any individual is performing whenever he or she plays a phonorecord embodying the performance or communicates the performance by turning on a receiving set.” This materially changes any causation analysis.

Second, although intent, foreseeability or knowledge are not required to establish strict liability, their presence certainly demonstrates a more proximate cause of the resulting harm. Some courts, like the Cablevision court, appear to see evidence of such factors as irrelevant, or even contrary, to direct liability. ⁴³See also Giganews: “Plaintiff ‘s allegations regarding Defendants’ knowledge of the pirated content on its servers do not salvage Plaintiff’s direct infringement claim. As the Netcom court pointed out, “knowledge” is not a required element of direct infringement (although it is a required element for contributory infringement).” However, it would seem that the fact that a service provider acted in a way that showed it intended the making of copies seems relevant to the question of whether it caused the making of the copies.

The claim that courts should analyze who “makes” a copy under proximate causation rather than volitional conduct is a starting point, not an end point. But it should provide greater clarity and results that are more congruent with common sense—results that should better adhere to copyright’s greater purpose.