Last week, a federal district court in New York issued a decision that, according to one report, “could change the way online publishing functions.” The court ruled that news websites that embedded a tweeted photograph of Patriots quarterback Tom Brady could infringe the photographer’s copyright. The ruling in Goldman v. Breitbart News Network (and also involving Time, Vox, and the Boston Globe) could present enormous problems for internet media practices.
But what is perhaps most surprising about this potentially internet-changing decision is that the court did not need to make the ruling at all. Indeed, the legal filings seem to suggest that the case was less about the photographer and more about the lawyers’ irritation with a legal test that arguably had no practical effect on the case.
At issue in the Goldman case is the common online practice of “embedding” content. When a website delivers an article to a reader, that article can contain Facebook posts, tweets, and all sorts of other material that may include text, photos, and videos. Yet the website itself may never store or send the actual content of those photos, posts, or tweets. Instead, the website sends to the reader’s computer instructions for locating the desired content on the servers of Facebook, Twitter, or another provider, and the reader’s computer is responsible for pulling up the content.
The technical workings of embedding are relevant to copyright law because there are two distinct types of copyright violation, both carrying the same penalties but requiring different elements of proof. First, you can commit what is known as “direct” copyright infringement by actually doing the unauthorized copying or distributing of the copyrighted work. But you can also help others perform those unauthorized acts in what is called “secondary” infringement. Secondary infringement requires proof of knowledge—a demonstration that the help given to others would lead to specific copyright violations. So, for example, distributing peer-to-peer file sharing software is not secondary infringement, even if you know generally that some people might use it for swapping pirated movies; it is only when you specifically know of and help individual movie pirates that secondary infringement occurs.
It may seem at first that Breitbart, Vox, and the other parties are clearly engaged in direct infringement, since the photo is displayed on the face of the article retrieved from the website. But that’s not correct, because of the way the technology works. Embedding cannot be direct infringement because the website never actually copies or distributes the copyrighted photo. Mere appearance of wrongdoing is not enough, generally speaking. If the law says that doing a thing is wrong, you actually have to do the thing to be liable. In this case, doing the thing would mean downloading the image from the tweet and then uploading it to Breitbart or Vox or Boston Globe servers to publish—not simply embedding. This is called the “server test.”
Embedding a copyrighted photo on a website, as in the Goldman case, could constitute secondary infringement, as the website transmits instructions helping the reader’s computer pull the photo. Generally, however, the embedding is still allowed because of other exceptions in copyright law, such as fair use for news reporting. But to deem embedding a direct infringement would cause problems in the many cases in which the website operator does not fully know what is ultimately embedded. Online advertisements, for example, are generally done through embedding, and the website operator often does not prescreen the ads. An embedded Facebook post can incorporate comments added after the fact. Even embedded tweets can end up including unexpected photos or text if the tweet contains a link to an article. In all these circumstances, the “knowledge” aspect of secondary copyright infringement protects websites that might otherwise unexpectedly find themselves copyright infringers due to the acts of others.
The “server test” came about more than a decade ago, from a federal court case from the U.S. Court of Appeals for the 9th Circuit in California. The case, Perfect 10 v. Amazon, decided that embedding is not direct copyright infringement. However, the “server test” specifically leaves open the possibility of secondary liability.
If the lawyers of the photographer in the Goldman case had wanted to, then, they probably could have pursued a ruling of secondary infringement against the news websites (so that the remaining dispute would be over other issues like news-reporting fair use). Yet the photographer’s lawyers consistently charged only direct infringement. Why? The legal filings seem to suggest that the photographer or, more likely, the lawyers simply wanted to challenge the server test.
Several legal filings reveal this grudge motivation. The complaint, which ordinarily would just contain “a short and plain statement” of relevant facts, strangely contains an extensive discussion of a conversation between the lawyers for each side. Even stranger, the conversation is not about the facts of the case, but rather hypothetical situations about photographs being displayed on televisions in bars, billboards of Trump ads, or baseball stadium displays—irrelevant examples, indeed, as they all involve physical ownership of the display screen, which news websites do not have.
The complaint then proceeds to hyperbolize over the server test, calling it a “contorted, out-of-context, highly technical, and exquisitely self-serving extrapolation” that serves as a “means for predatory publishers to avoid paying license fees for their increasingly rampant appropriation of copyrighted content” without “any other positive social purpose.” And later briefing on behalf of the photographer continues to criticize the server test as a “sweeping industry-wide immunity for all manner of copyright theft.” (Disclosure: I helped to prepare a friend-of-the-court brief opposing that position.)
Those strong words would be justified if not for the fact that the server test does serve the positive social purpose noted above—preventing unexpected liability from unpredictable embedding—and if the lawyers had not ignored the plain fact that they had a perfectly reasonable alternative way of bringing their case: through secondary infringement.
Unfortunately, the federal district court bought those arguments and disregarded the server test (as it is somewhat allowed to do, because a New York district court is not bound to follow the 9th Circuit). More unfortunately, there may be no simple avenue to have the decision corrected. The Goldman case likely can only be appealed if the court ultimately decides against the news websites. There are several other ways the websites could win, and the photographer might drop or settle the case having gotten the one ruling he apparently wanted. As a result, this problematic decision, and the questions it raises for online advertising and embedding practices, could remain on the books for years to come.