On August 11, 2010, media giant Viacom filed notice with the Second Circuit that it would be seeking an appeal in its legal battle against YouTube. The Southern District Court of New York essentially threw out the case, granting summary judgment to YouTube on June 23. Viacom originally sued the video site in 2007, claiming that the site did not do enough to prevent users from uploading infringing videos. The stakes are high – Viacom alleges $1 billion in damages, and Google has already spent $100 million to defend. More importantly, the case may have ramifications for many sites that rely on users for content.
What is this case all about? Some have suggested that Viacom is trying to “get back” at YouTube after it failed in a bid to buy the video site. Or even more sinisterly, maybe the case is an example of the entertainment industry’s shift toward propping up failing business models by extracting settlement money from legal threats. Exploring those contentions may have merit, but that’s not what this site is for. Instead, I’m more interested in examining the legal principles at work to get a better grasp at what’s going on inside the courtroom.
The case boils down to this: what obligation does a host like YouTube have to keep infringing materials off their site? To understand this a little better, let’s first take a closer look at secondary liability and internet safe harbors.
Qui facit per alium facit per se.1
By default, we are not responsible for someone else’s actions. There are circumstances however where the law does hold one person responsible for the actions of another. In criminal law, for example, someone aiding and abetting a criminal can be charged with the same offense. In civil law, the concept is sometimes referred to as secondary liability.2
It makes sense. If you have some sort of control over another’s actions and you benefit from their wrongdoing, or if you help someone knowing they will use your help to commit a wrong, you shouldn’t be let off the hook. Secondary liability is also a form of risk allocation. Holding someone responsible for the acts of those in their control encourages safety and fair business practices. Finally, secondary liability addresses concerns of obtaining redress when someone is wronged. It is more efficient to sue a single employer rather than multiple employees, and the person running the show often has deeper pockets than those under his control.
Secondary liability is well-established in many areas of the law. Let’s turn now to how courts have used it in the copyright context.
Secondary Liability in Copyright
Courts have applied two separate theories of secondary liability to copyright cases: vicarious liability and contributory infringement.3 Vicarious liability includes the concept of respondeat superior, used to hold employers responsible for the acts of their employees, and requires that one benefits from conduct within one’s right or ability to control. Contributory infringement requires providing the means to engage in infringement with knowledge that infringement will occur.
One point of confusion: it’s easy to conflate secondary liability with direct infringement, but important to keep the two ideas separate.4 Direct infringement requires an unauthorized reproduction, distribution, or public performance. Secondary liability is liability for indirect infringement. The person doing the actual reproduction or distribution is directly liable; anyone who is not engaged in those acts but is contributing to – or controlling and benefiting from – that direct infringement is indirectly liable.
While vicarious liability is fairly easy to find in older copyright cases5, contributory infringement was relatively rare before the internet became popular. It became easy for anyone connected to transmit all kinds of data, including copyrighted content – and transmit they did, first by the hundreds, than by the thousands, than by the millions (often anonymously). The transmission of this content was made possible by several different parties: ISPs, web hosts, BBS operators, etc. Which of these parties should also be held responsible for infringement by internet users?
Needless to say, courts struggled with applying traditional legal doctrines to new, unfamiliar technologies. Concerned about the effects liability would have on the growth of an important communications network and the problem of having inconsistent law throughout the country, Congress passed the Online Copyright Infringement Liability Limitation Act (OCILLA) as part of the Digital Millennium Copyright Act of 1998 (DMCA).6
OCILLA provides online service providers7 with a conditional safe-harbor from secondary liability. For anything stored on a service provider’s system or network at the direction of a user – e.g., a video uploaded to YouTube – service providers can qualify for this safe-harbor if they do not have knowledge of the specific infringing act, do not benefit directly from the infringement, and remove the infringing material when they receive notification from a content owner. This last aspect is known as a “notice-and-takedown” procedure.
So, continuing with the YouTube example, a user who uploads a television clip to YouTube has directly infringed – he has reproduced and distributed a copyrighted work without authorization from the owner. YouTube would ordinarily be indirectly liable for this infringement, either because it has provided the means for the user to infringe or because it has the right and ability to control what is presented on its site and benefits from the traffic its content attracts. But under OCILLA, it is shielded from this potential liability if it removes the video once it either has knowledge that it is infringing or when the content owner sends a DMCA notice.
Viacom v. YouTube
The heart of the case between Viacom and YouTube hinged on interpretation of the “knowledge” component of the safe-harbor. The full language of the statute states that a service provider only qualifies if it
(i) does not have actual knowledge that the material or an activity using the material on the system or network is infringing; (ii) in the absence of such actual knowledge, is not aware of facts or circumstances from which infringing activity is apparent; or (iii) upon obtaining such knowledge or awareness, acts expeditiously to remove, or disable access to, the material.
Viacom alleged that YouTube was acting in “willful blindness“. It was well aware that large amounts of content on its site was infringing but chose to ignore it – welcoming it, even – unless it received notification from the content owner. Essentially, Viacom argued, this awareness of infringement in general constituted the “actual knowledge” that triggers YouTube’s obligation to remove the material without receiving notice from the content owner.
The court sided instead with YouTube, and held that actual knowledge meant “actual or constructive knowledge of specific and identifiable infringements of individual items.” “General knowledge that infringement is ‘ubiquitous’ does not impose a duty on the service provider to monitor or search its service for infringements.”
That leaves the second prong of the knowledge component – a lack of awareness of “facts or circumstances from which infringing activity is apparent.” This is sometimes called “red flag infringement,” as in the facts and circumstances surrounding the activity should raise a red flag that infringement is occurring.
After reading the case though, I couldn’t figure out what, if anything, constitutes a “red flag”. Several others have pointed out that courts have largely written out the “red flag” prong. Here, the court wholesale adopted reasoning from Perfect 10 v. CCBill, a 9th Circuit case dealing with “red flag” trademark infringement. Copyrights and Campaigns author Ben Sheffner notes how unconvincing the reasoning of that case was:
In other words, under Ninth Circuit precedent, having material identified by its poster as “illegal” and “stolen” is not a red flag that infringing activity is taking place. One is left to wonder whether the panel would have ruled the same way had actual red flags been waved in the defendants’ faces.
One of the basic rules of interpreting a law is to avoid rendering any part superfluous.8 Courts assume lawmakers had a reason for wording a statute the way they did; any interpretation that vitiates a word, phrase, or entire provision subverts the lawmaking process.9
So the question is, what exactly would count as a “red flag”? It would have to be something less than specific knowledge that a particular work is infringing, but something more than general knowledge of infringing activity, including material specifically identified as being “illegal” or “stolen.” It’s difficult to think of anything that would fall in that middle ground, and it is possible that this interpretation of “red flags” will be revisited on appeal.
Why This Matters
This ruling has little effect on YouTube. Its policies have changed substantially since the suit was filed; today, many content providers have few qualms with YouTube’s approach to infringement, and the site continues to develop technological solutions to police infringement more effectively, especially its Content ID program. In addition, the pending appeal means that the lower court’s holding has little effect.
But the issues at play bring up a larger question: how much burden should be placed on hosts for policing infringement when they encourage user-generated content to drive traffic to their sites and, ultimately, sell advertising (or make money off the traffic in some other way)? While the DMCA has given content providers a mechanism to remove infringing material that is far cheaper and faster than going to court, the scale of the internet and ease of reproducing and distributing content means that creators still must focus an inordinate amount of time on managing infringement. YouTube is just one of many sites.10 If we talk about secondary liability in the context of risk allocation, should sites like YouTube – which receive a direct benefit by allowing user-generated content and have a greater ability to monitor what is uploaded on the back-end than multiple content creators have on the front-end – have more of an obligation to minimize direct infringement?
We’ll have to wait and see.
- “He who acts through another does the act himself.” [↩]
- It is also sometimes called third-party liability, a linguistic oddity that makes me chuckle. [↩]
- A third type of secondary liability, inducement, was introduced in MGM v. Grokster, but it is still unclear whether the Supreme Court meant to create a new, discrete form of secondary liability or whether they were merely rephrasing the concept of contributory infringement. [↩]
- Some courts have even fallen into this trap – see Patry on Copyright § 21:40 [↩]
- Many cases involve unauthorized public performances of music by bands hired by dance halls [↩]
- OCILLA was codified as 17 U.S.C. § 512. [↩]
- “Online service providers” encompasses any company or service that transmits or communicates information between users and includes internet service providers and web sites. [↩]
- CRS Report for Congress, Statutory Interpretation: General Principles and Recent Trends. [↩]
- This is different from a court striking down a law on constitutional grounds. If the constitutionality of a law is not being challenged, courts generally don’t have the power to change the meaning of laws to whatever suits them. [↩]
- Viacom notes that monitoring YouTube alone is time-consuming. [↩]