Broadband Isn’t a “Draw” for Infringement, but What About Substantial Non-infringing Uses?

Back in late 2014, two of Rightscorp’s clients, BMG Music and Round Hill Music, sued the cable operator and internet-service provider, Cox Communications, for copyright infringement on grounds that Cox was liable for its users’ sharing of the plaintiffs’ copyrighted musical compositions using the BitTorrent protocol.

What’s Right for Rightscorp

Although it wasn’t a party, the case was crucial to Rightscorp. Rightscorp is in the business of investigating the sharing of copyrighted work over BitTorrent protocol1 and obtaining modest settlement from the BitTorrent users. For example, if Rightscorp thought you had shared “Bad Blood” using BitTorrent, it would send you a settlement demand of, say, $500—or some figure that’s low enough for you to afford but not high enough to be worth fighting over. It’s a low-return–high-volume business. And for it to work, Rightscorp needed to get as many settlement demands to users as possible.

But to do that, Rightscorp needed the cooperation of ISPs. That’s because Rightscorp doesn’t actually know who the user is. It just knows the user’s IP address at the time of the alleged file-sharing. Only the user’s ISP knows which of its users was using a particular IP address at a particular time. And the ISPs have little incentive to just hand that information over to Rightscorp. These are, after all, customers; and selling out customers is generally bad for business.

Rightscorp thought it found a way to force ISPs to play ball. It would send to the ISP DMCA takedown notifications that also included a special message to the user: pay up or else. The point wasn’t to get get the infringing material removed. An ISP can’t do that.2 In fact, DMCA takedown notifications are, strictly speaking, legally meaningless when sent to ISPs. But ISPs do have an obligation to develop and “reasonably implement” a “repeat infringer policy”—if they want to be protected by the DMCA safe harbor. And a requirement for that is to have some kind of open communications channel for learning about your users’ infringing activities.

The point of those DMCA takedown notifications was extract settlements from the ISPs’ customers. Rightscorp had it in its head that the ISPs would be legally obligated to forward on the notifications to the customer. I haven’t the slightest idea why Rightscorp felt that way. The only reason a takedown notification would be forwarded to a customer is if (1) the recipient felt like it, or (2) the recipient wishes to implement a counter-notification system, which is also effectively voluntary.3 Since ISPs don’t store information, reason (2) doesn’t apply. There’s nothing in the statutory language for the repeat infringer policy that even remotely suggests notifications must be forwarded to anyone. All that is required is that the ISP respond to the information in a way that reasonably implements a repeat infringer policy.

Sailing in Dangerous Waters Outside the (DMCA) Safe Harbor

The DMCA safe harbor is crucial to any ISP. That’s because ISPs have millions of customers, any number of whom are infringing copyright at any time. Without the protection, ISPs would be potentially4 liable for those users’ infringing activities at any time. Maybe not enough to bankrupt the ISP—which are usually extremely large and wealthy—but enough to put a noticeable dent in its profits.

Fortunately, it’s dead easy for ISPs to avail themselves of the DMCA safe harbor. Unlike companies that store information at their customer’s request, like YouTube and Dropbox, the requirement for ISPs consists almost entirely of stuff ISPs normally do anyway. The only exception is the repeat infringer policy. But even that isn’t too hard to comply with. While we don’t know too much about the precise requirements—Congress couldn’t have been much more vague in its wording—we know it at least requires a means to obtain information about infringement and a policy that can realistically result in the termination of repeat infringers.

As I recounted last time, Cox managed to screw this up by only pretending to terminate even inveterate repeat infringers. Shorn of its DMCA safe harbor, it was in some peril.

But not absolute peril. The DMCA safe harbor is a defense. The plaintiffs still had to prove that Cox was liable for its users’ infringing activities. To do that, the plaintiffs needed to prove that Cox either contributed to that activity, or was vicariously liable for that activity.

Loose in the Secondary

I was originally skeptical that the plaintiffs could prove either.. However, we’ve learned a few things about Cox that have strengthened the plaintiff’s case, though they still had work to do.

To be clear, Cox was never accused of directly infringing the plaintiffs’ copyrights; i.e., no one accused Cox of reproducing, performing or distributing the plaintiffs’ copyrighted materials. Rather, the claims were for secondary liability.

There are two main flavors of secondary copyright infringement: contributory and vicarious. Contributory infringement requires proof of (a) direct infringement by someone else, (b) knowledge of the direct infringement, and (c) materially assisting the direct infringement, such as by providing access to the Internet, over which the illegal file sharing takes place.5 Vicarious infringement requires proof of (a) direct infringement by someone else, (b) the right and ability to control the activity that’s causing the direct infringement, and (c) direct financial benefit from the infringing activity.

As you might have heard, a jury has recently found Cox liable for contributory copyright infringement (but not for vicarious), and assessed statutory damages in the amount of $25 million. Considering that Cox’s revenues are about $17 billion, I’m not sure this is a very big deal for Cox. And it’s probably not going to be a sea-change for copyright enforcement, either.

It’s OK if You Don’t Want to Read my Article on the Subject

Both legal theories required proof of direct infringement. I was originally skeptical the plaintiffs could do that, but the jury must have found direct infringement. Rightscorp couldn’t see the files in transmission, so it couldn’t know which files had been distributed. What it could see was the list of files that a user had made available for others to torrent. Most courts have held that “making available” isn’t, without more, an act of copyright infringement. And the court in this case was no different.6 The court did not, however, dismiss the plaintiffs’ claims, for a couple of reasons.

First, it turns out that Rightscorp itself had torrented some of the files, and these torrents counted as distributions.7 Second, the court held that the sheer number of files that were made available for torrenting was itself evidence that jury could use to infer distributions.8 But this raises an interesting problem. It’s one thing to say that some of the available files were distributed and, therefore, there is direct copyright infringement. It’s quite another to say how many of those files were distributed—or, more precisely, how many works were distributed. That’s important to know because statutory damages are based on how many works were infringed9 If there were, say, 10,000 works made available, of which 2,500 were owned by the plaintiff, how does the jury figure statutory damages? I’ll return to this conundrum later.

Draw! Or Not

The jury found there was no vicarious infringement. I had been especially skeptical that the plaintiffs could prove that Cox had benefited financially from the infringing activity. Normally, when you charge everyone the same for your service, and some use the service to infringe but others don’t, you aren’t considered to have benefited financially from the infringement. That’s because you received the same money regardless of what your customers were doing. The exception to the rule, however, is if the ability to perform the infringing activity is a “draw” to use your service. I.e., if people couldn’t infringe, many of them just wouldn’t bother using your service. For example, both Napster and Grokster have been found to benefit financially from infringing activity because they were both specially designed to facilitate file sharing and nearly all of their users were sharing files illegally.

I didn’t think Cox was a “draw” for infringing activity the same way Napster and Grokster were. Cox’s service is just broadband. Customers use broadband for all kinds of things, the vast majority of which is perfectly legitimate. Cox wasn’t any better set up for torrenting than any other service. More important, to my view, no one really has a choice but use Cox (or possibly a substantially similar competitor) if they want access to the internet. The plaintiffs introduced evidence of a survey showing that 16% of Cox’s users illegally share music, and of those 16%, 70% subscribed to Cox for that reason.

When Cox tried to dismiss the vicarious liability claim at summary judgment, the court ruled that there was enough evidence to go to the jury. The court held that, because of the survey evidence, a jury could find that illegal file sharing was a reason to subscribe to Cox’s service. Evidently, the jury decided against the draw theory.

The court also held that Cox’s contractual ability to terminate users for (among other things) copyright infringement was enough for a jury to conclude that Cox had the right and ability to control the users’ infringing activity.10 Again, I was skeptical at first, because there was a good deal of legal authority holding that such gatekeeping wasn’t the same thing as “control,” which is more nuanced. The court rejected that view (which was its right—none of that authority was binding on this court), and so that issue went to the jury. I’m guessing the jury didn’t even reach it during deliberations, because the “draw” theory was a more straightforward issue.

Just Here to Contribute

The jury did find there was contributory infringement. There was plenty of evidence that Cox’s services were a constituent part of the acts of direct infringement. Also, in light of the evidence that Cox knew of specific infringing activity, it’s not terribly surprising that the jury might find the knowledge requirement had been met. The harder question what acts of infringement did Cox have knowledge of? Did the jury find that Cox necessarily had the requisite knowledge as to all of the plaintiffs’ works? Or just some of them? It’s hard to tell.

I was nevertheless skeptical—and, honestly, remain skeptical—that the plaintiffs could prove contributory infringement because there is a wrinkle to the analysis when the claim is based on some product or service that is being offered to the public. Under the Sony/Betamax and Grokster decisions, there is no contributory infringement if that product or service has “substantial non-infringing uses,” unless it could be shown that Cox induced the direct infringement. In Sony, the device at issue was a VCR, and the Supreme Court held that, because VCRs could be used to “time-shift”11. But in Grokster, the Supreme Court cabined any expansive readings of Sony. Although Grokster provided a service that, in theory, had lots of non-infringing uses (that no one took any advantage of, natch), it lost anyway because it went out of its way to help its users commit copyright infringement—i.e., it induced infringement. But Grokster otherwise confirmed Sony general rule about substantial non-infringing uses.

Cox, it seems to me, is a lot more like Sony than like Grokster. The point of Grokster was infringement. But Cox’s service is just a vanilla broadband internet access. People stream movies on Netflix using Cox a lot more than they torrent files.

But Who Instructs the Instructors Themselves?

Why didn’t the jury see it this way? The answer might lie in the jury instructions. The parties fought hard over how to describe the Sony and Grokster holdings. Cox described it pretty much the way I just did. The plaintiffs contended that Sony and Grokster had no bearing whatsoever on contributory infringement. Sony is not cited once in the plaintiffs’ proposed instructions, and Grokster is cited only in connection with other issues.

The actual jury instructions are not part of the docket. Perhaps the court will file them at some point, but right now, if you want to know how the court dealt with Sony and Grokster, you’d need to get ahold of the transcript for the trial transcript (where the judge patiently reads all of the jury instructions to the jury12).

Math Is Hard, and Not Totally Necessary

Where in the world did the jury come up with $25,000,000? We don’t know that at all. Because the jury found that Cox’s contributory infringement was “willful” (I’m guess it’s because of its deceptive use of its termination policy), the jury could have awarded up to $150,000 per work infringed (and as little as $750). We also know from the plaintiffs’ proposed jury instructions that the plaintiffs asserted a total of 1397 works infringed. That leaves a range of $750 (if the jury found only one work was infringed, and awarded the smallest amount possible) to just over $209.5 million.13

If the jury found all of those works had been infringed, then it awarded $17,895.49 per work, on average. But it’s possible the jury found only some of the works were infringed but awarded a higher average amount per work. Remember, the plaintiffs didn’t have direct proof that all 1397 works were actually shared, only that they were “made available.” For example, the jury might have found 250 works were infringed and awarded $100,000 per work. The jury isn’t obligated to award the same amount for every work, or even to award an amount for every work. So long as the jury’s award is somewhere in the permissible range—between $750 and $209.5 million—it’s OK. All we can deduce for certain is that the jury must have found at least 167 works were infringed.

But how was the jury able to determine how many works were infringed? The only evidence before it was (1) evidence that Rightscorp had downloaded a certain number of those 1397 works, and (2) the rest were merely “made available.” The ones that were actually downloaded are a known quantity, but how is the jury to know how man of the remaining 1397 works were downloaded? Does it just guess that it was, say, one third? Does it assume they all were? Was there expert testimony explaining what proportion of a file made available for torrenting was actually torrented?

More light will be shed when the parties file post-judgment motions (there’s no schedule for those, yet). Given my interest in the whole “making available” thing, I’ll be most interested in how the judge treats the evidence that bridges the gap from making available to actual distribution. I suspect the judge will also have to wrestle with how to implement Sony and Grokster. Do those cases really require a lack of substantial non-infringing uses and inducement, and if so, was there really evidence to support the jury’s verdict on that issue?

Live and Don’t Learn, That’s Us!

In the end, however, this case was a whole lot about nothing. We learned almost nothing about the requirement for repeat infringer policies, except that you shouldn’t implement it in a deceptive way (and document your deception), which, you know, should’ve been obvious from the start. We learned that vanilla broadband services are probably not “draws.” We learned that the going rate for infringement is probably about $20,000 to $25,000 per work, which is a lot for you and me but peanuts for large ISPs. Other than being more careful with their repeat infringement policies, ISPs aren’t likely to adjust their behavior, and wide-scale piracy isn’t likely to be much affected.

Thanks for reading!