Server Test Out, Transmission Test In

If you’ve followed United States copyright law in past 20 years, you’re undoubtedly already familiar with the Server Test.

Established in May 2007 by the Ninth Circuit Court of Appeals in the case of Perfect 10 v. Google. The test has been important and controversial in equal measure. In short, it broadly legalized the embedding of copyrighted works so long as the allegedly infringing site is not hosting a copy of the work on its server.

Since then, the server test has come under repeated attack. Other circuits have rejected it, and it seemed like the test might be on its way to a Supreme Court review.

Now, a new ruling in the Fifth Circuit Court of Appeals, theoretically, poses a new threat to the server test. The court ruled that the Ninth Circuit’s decision in the Perfect 10 case was deeply flawed, relying on the fixation of an allegedly infringing work, rather than the transmission of the work.

However, in scrapping and rewriting the rule, the Fifth Circuit actually landed pretty much in the exact same place as the Ninth Circuit. Though it took a very different path to get there, the outcome is very similar.

As such, it’s worth taking a moment to examine the Fifth Circuit’s decision, how it differs from the server test and why, for most people, it’s functionally the same thing.

Background on the Cases

The Perfect 10 case pitted the adult entertainment company Perfect 10 against Google. Perfect 10 filed the lawsuit alleging that Google infringed their copyrights by displaying full-size images as part of their image search results. However, the Ninth Circuit found that Google was not liable for copyright infringement since the images were hosted on third-party servers, not Google’s.

Tech groups largely cheered the decision, saying that it provided much-needed clarity on who is liable for copyright infringement. Artists, however, complained that this opened the door to widespread embedding of infringing material without liability.

Other courts, including a district court in the 10th Circuit, have outright rejected the server test and found that embedding unauthorized content can be an infringement. That said, though the server test isn’t binding outside of the Ninth Circuit, it is still the leading guidance on the issue.

That is, until now (possibly).

That circuit recently heard an interlocutory appeal in the case of Emmerich Newspapers v. Particle Media (NewsBreak). Emmerich newspaper operates a series of local newspapers in the Mississippi area. Particle Media is best known for NewsBreak, its news aggregation service.

Emmerich sued Particle Media for copyright infringement, claiming that NewsBreak would display their full articles in a framed result with NewsBreak’s logo and ads. In addition to copyright infringement, Emmerich claimed that NewsBreak was removing copyright management information (CMI) from their articles by hiding their URLs.

The district court applied the server test and found that Particle Media was not liable for copyright infringement. This prompted Emmerich to appeal, citing two questions:

1. Whether the “server test” is the proper standard for evaluating infringement on the copyright holder’s exclusive display right?
2. Whether a URL can be considered CMI and, if so, whether removal of the URL is a violation of that CMI?

The court outright rejected the server test. It found that the Ninth Circuit’s decision was deeply flawed as it relied on a file being “fixed” when there is no need for a fixed copy to transmit it. Instead, it looked at who transmitted the work, not where it was hosted, and found that, in this case, it was the same entity, Emmerich.

Though the Fifth Circuit did not rule on the facts of this particular case, the way it presented the transmission test seems to indicate that Emmerich had the right and ability to prevent the embedding of their articles and transmitted them anyway. This, most likely, does not bode well for Emmerich when the case goes back to the district court.

The court left open the CMI question, finding that URLs can be CMI but only if they convey information about the work that goes beyond just an address. It did not rule if that bar was met in this case.

The appeals court then remanded to the lower court for factual determinations based on its answers to these questions.

The New Test, Same as the Old One

To be clear, the Fifth Circuit’s decision leaves a LOT of questions unanswered. As Aaron Moss a Copyright Lately put it, “Permission to display a copyrighted work, a server’s technological willingness to answer an embed request, and an implied license allowing a particular third party to embed are three different things. Emmerich moves among all of them without completely sorting out how they fit together.”

It also didn’t determine if the transmission that took place in this case was authorized or permitted. That was left for the lower court to decide on remand. So, while we have this new test, we have no idea how it will actually be applied.

But there is one truth that is very clear: On the internet, the person who is storing the work is probably the same person who is transmitting it.

There’s very little functional difference between “The articles were stored on Emmerich’s servers” and “The articles were transmitted by Emmerich’s servers.” Granted, the transmission test does lend some more complexity to to the standard, but for most cases of unwanted embedding, it’s up to the copyright holder to prevent the embedding rather than the other party to simply not do it.

The Fifth Circuit also doesn’t address what happens if the copyright holder can’t prevent the embedding. What happens if it’s on a service where there is no way to turn off embedding? What if it’s a pirated copy of the work on a third-party site? The Fifth Circuit doesn’t address any of these issues.

This decision ultimately creates more uncertainty than it resolves. Now there are two similar, but slightly different, tests and one of them is only somewhat defined. To make matters worse, it might make getting that clarity more difficult.

As Moss pointed out, by creating a different test with largely the same outcome, the Fifth Circuit has avoided the kind of circuit split that the Supreme Court is likely to take up.

So there the waters are significantly more muddied than they were before and we’re farther away from a Supreme Court decision that could provide real clarity.

It’s a rare time where, no matter how you feel about the server test itself, this feels like a step backwards, even if the logic is, ultimately, more sound.

Bottom Line

Regardless of what one thinks of the outcome, the server test was always flawed and the litany of cases that have gone against it have proven that. The fixation requirement of the test was deeply flawed and only of limited use in the modern internet.

The transmission test, from a legal perspective, is much better. It eliminates the need for a fixed copy and focuses on the actual transmission of the work.

However, we’ve had nearly 20 years with the server test. For all its flaws, it’s the devil we know. Now we have a new test that is similar, but not the same. Now we have the devil we don’t know. All that we know is that it looks, sounds and feels very similar to the first one.

To be clear, I’ve always been a critic of the server test. I always felt hat it fundamentally misunderstood how the internet works and what users experience. I thought, following the Aereo decision in June 2014, that the server test was on its way out the door. But it never happened.

This new transmission test does a better job of addressing the seeming divide between the server test and the Aereo decision. There’s a reason why the Fifth Circuit’s decision cites the Aereo decision 11 times.

But creating a new test with so many unanswered questions injects a great deal of uncertainty into the equation. This case will be one to watch as it goes back to the lower court. But as Moss said in his post, “Whether the new test ever produces a different outcome will be somebody else’s lawsuit.”

That’s the grim truth, we won’t know what this new test means for litigants until more cases come along and the actual boundaries get drawn.

That could take literal decades.