Annemarie Bridy – Three Notice Failures in Copyright Law
Challenge of making online copyright enforcement meaningful. Notice is intimately connected to justice: usually required (except with strict liability). Failures to appreciate its necessity/failures to appreciate its sufficiency.
Uncertain: the construct of red flag knowledge in the DMCA. No notice: lack of notice in seizure of domain names under criminal © law. Naked notice: in preliminary injunctions against nonparty service providers in pirate site cases.
Purpose of DMCA safe harbors was certainty for ISPs in the face of evolving secondary liability in cyberspace. In practice, years of litigation have clarified the test to be part subjective/part objective. Does the service provider have subjective knowledge of facts that would give objective knowledge of infringement to a reasonable person. Red-flag knowledge as a safe harbor disqualifier can preclude summary judgment; undermines certainty of notice and takedown framework; creates risk for ISPs and encourages proactive takedowns outside the DMCA checks and balances.
The fix: make actionable knowledge under the DMCA synonymous with the knowledge that comes from receipt of compliant knowledge.
Notice failure #2: PRO-IP Act added asset seizures and forfeitures to remedies available in criminal © cases. Get an in rem arrest warrant for the domain name, w/o notice to registrant—civil forfeiture even if gov’t never charges or proves an underlying crime. 2010-2013, FBI/DOJ/ICE went after 1700 domain names. Tend to get active around the holidays. Megaupload, 2012, actually accompanied by an indictment. Sept. 2015: ShareBeast.
Notice failure under both 1st and 5th Amendment, which requires notice and opportunity to be heard before property seizure, unless exigent circumstances exist where property could be moved before seizure. Domain names aren’t movable property, though the content to which they provide access is. Seizing a domain name does nothing to prevent underlying content from being moved. First Amendment also requires notice and opportunity to be heard before a seizure of expressive property—if the gov’ts reason for seizing expressive property is to take it out of circulation (as opposed to evidentiary). Domain names are expressive property b/c they provide a gateway to vast repositories of speech.
The fix: apply the correct legal standard and provide notice and an opportunity to be heard; more than probable cause. Pending legislation would increase the burden of proof for all covered crimes, including ©, though it may not go anywhere.
Naked notice: Site-blocking TROs and PIs purporting to enjoin search engines, payment processors, ad networks, domain name registrars and registry operators if they have notice of the injunction under the All Writs Act. These are improper. Due process, separation of powers issues. Only nonparties so closely identified w/the defendant that their interests can be considered to be represented can be bound—notice and a finding that the nonparty was in active concert w/defendant in the illegal conduct—aiding and abetting. Arms’ length provision of tech services to infringers is not active concert.
Michael Fricklas, GC of Viacom: Most of what we do doesn’t happen in cases—negotiated outcomes in disputes; our decisions on what to put on the air. Some failures are litigated, and others aren’t. Every area of the law has some level of uncertainty.
Copyright’s strength is that you don’t need gov’t support—a plebiscite every day where people decide what to watch and what to pay. In that context, courts have been struggling with importance of certainty, but in the cases Bridy addresses, what we’re doing day to day doesn’t constitute a notice problem at all. The targets of our enforcement efforts have no doubt about infringement. Even injunctions against nonparties aren’t serious b/c the courts are asking only for an administrative act. Not finding that nonparties have violated copyright law (though they might be if they continue to provide services after an adjudication of infringement).
Last year we took down 1.8 billion pieces of content, with 100 million IP addresses. If we made a mistake you can email us and don’t need to wait for counternotification on YouTube. We can’t do fair use analysis at this volume. So we took simpler rule: we care mostly about entire/substantial part of content posted w/o alteration. Not interested in mashups. Viacom uses fair use every day: Jon Stewart and Stephen Colbert—we get sued a lot.
Certainty isn’t the most important value embodied in the DMCA. Also: balance interests of © owners and ISPs by creating strong incentives for them to cooperate in dealing with infringement. Red flag knowledge is a boundary case. We spent way too much money on YouTube case, but the court was basically thinking about who were the good guys and bad guys. Court didn’t understand problems with notice and takedown—no one has to notify us where they’re posting; people can do all sorts of things to hide from automatic detection; ISPs will limit how much you can search, and will hide the source of the infringement for a fee. Senate anticipated some of these issues (heh), striking the right balance with red flags. ISPs not required to make discriminating judgments, but seeing an obviously pirate site may be all that’s needed. Doesn’t mean they’re liable, but that the safe harbor no longer applies. Actual knowledge will also disqualify you. The cases have said that vicarious liability still exists if there’s intent to infringe.
Notice cases in SCt were about due process, not about boundary of copyright—SCt says due process is flexible and requires the protections demanded by the situation. Balances private interests w/administrative costs imposed on gov’ts interest and risk and probable value of additional safeguards. 1700 seizures of domain names, but only 2 problems. Don’t forget the obvious cases, where people have full-fare information about what they’re doing. Megaupload: indictment says they knew they were infringing. Seizing a bank account from a bank doesn’t require that the bank was involved in a conspiracy. (Ok, that wins for most misleading analogy, since the seizure of the domain names (bank account) is precisely done b/c the “bank” (registrant) is a wrongdoer.)
Joseph Liu – Notice Failure, Fair Use, and the Limits of Property: Literature discusses fair use uncertainty (and counterliterature, including from Pam Samuelson, discusses how that may be overclaimed). Questions about validity of © are typically not at issue b/c of low © standard; issues of ownership can be troublesome given length of © and transfers, but not generally implicated by fair use disputes; the one area where there is potential notice failure is the Q of clearly delineating the boundaries of the entitlement. Even here, one might ask whether notice failure adds much to our discussion, b/c issue is not so much metes and bounds of particular © as with patent, but rather w/the structure of © law itself. Uncertainty may be only that caused by standard and not rule.
Core is relatively well-defined: generally can’t make large numbers of full copies and sell them w/o permission, or make major motion picture based on film. But at the boundaries there is uncertainty.
Information burden isn’t evenly distributed. Larger/more sophisticated parties can mitigate more easily by seeking custom legal advice or by bearing transaction cost of licensing. Unsophisticated parties have trouble getting either. Tech changes have exacerbated disparities because fair use used to be less relevant to the general public—personal use and other ways of engaging with © works and more directly subject to ambiguities. Core and periphery have shifted.
Dealing w/uncertainty: more best practices? Again, less available to the public. Ask market/tech to cure—costless licensing/technological fences that specify what people can do. But © owners’ troubling tendency to fence in what they don’t own weighs against that.
What if instead of trying to clarify the boundaries, particularly for less sophisticated parties, we considered alternative frameworks for liability? Ask whether fair user had adequate notice of the boundaries of the entitlement—for an unsophisticated party the answer will often be no. © for general public bears little resemblance to property as clear rights with robust transactional markets. Confront notice failure directly: try to conform standards to people’s expectations/negligence or tort approach.
Jessica Silbey: Sidestepping incentive talk and focusing on the way that © affects us all; notice doesn’t necessarily work for us but on us. How does the fair user understand the scope of property rights? How does the fair user understand what © is for? The second question gets to a similar place, but framing can teach other lessons. From property/markets to culture/fair, creative practices—creator in the community. Many creators she talked to have high tolerance for copying because everyone borrows; they demand a high standard for originality, reciprocally. If they find unfair copying, they want attribution and proportional remuneration—profit-sharing or even nominal, dignitary fee. Do they think about fuzzy boundaries? No, but every day © users tolerate a less than perfectly understood system, largely by intuiting rules that are misaligned with the © system. They believe that infringement and damages incorporate reasonability determinations. It would make sense to base fair use on reasonability determinations. Copyright owners do not have rights where fair users do; fair use is the baseline.
Samuelson: Molly van Houweling has a similar project—also, what do you think of Tim Wu’s “tolerated use”?
Liu: van Houweling’s work on new servitudes is important for tracing rights/impact on unsophisticated parties. Thinking more expressly about distributional impacts—empowers certain people to do more, which may be an important value independently. On tolerated uses: still grappling with that. One response to this is that maybe we don’t need to worry about unsophisticated users b/c we deal with that through underenforcement/no one sues individuals. I resist that b/c it’s so clearly second-best solution to problems w/actual standard; fuzziness and underenforcement might not match up. Custom and toleration might affect the fair use line.
Orly Lobel: continued discussion of distributional effects from previous panel. Sophisticated/unsophisticated creators; developed/less developed countries; employer/employee. It’s not just repeat players and institutional planners shaping the substance of the law and notice can counter that; they also will be shaping notice itself/what notice is. When we consider over/underdeterrence, consider the expressive function of law in general. The psychological effects of the FBI warning.
Liu: copyfraud/addressing misinformation might be part of a solution. Copyright Office could play an educational rule on what people can and can’t do. Copyright ombudsperson: role to look after the public interest.
Litman: Implicit in all 4 talks was that their either should be or already is an unacknowledged mens rea requirement for © infringement. If that’s right, what ought it to be.
Fricklas: take into account there’s misinformation on both sides—users think they can post a whole work if they comment on it. Intent may apply more to intermediaries than initial infringers. Suppose my search finds a 100% complete match—how do I do a mens rea analysis w/o a complete collapse of the system? Mens rea can be important, but sometimes res ipsa loquitor.
Bridy: we want a higher standard for secondary liability—should it be higher for direct infringement as well? It’s not always so obvious what’s obvious to a reasonable person; the edge cases are what lead to the super-expensive litigation. YouTube case settled w/still opened questions; Veoh got litigated into bankruptcy; Vimeo might give us some more info if the Second Circuit ever decides it. I favor more certainty—rules over standards. So that would drive me to higher mens rea.
Silbey: intent to copy v. intent to harm—unreasonable position to say that people should have to not intend to copy to avoid infringing; most copiers intend to copy. I would want a harm requirement instead. People who think they’re not doing anything wrong are applying a harm standard.
Liu: on effective mens rea standard: my sense is when it comes to whether bad faith is an element, Beebe found it generally wasn’t. Should be: tricky. Not willing to go that far. For unsophisticated, modify standards/remedies?
Bridy: sophisticated = proxy for knowledge.
RT comments on Bridy’s paper: red flag as unworkably uncertain. Strikes me as an overstatement, given litigation outcomes and the current persistence of competition. A slightly different diagnosis, and one that might point to competition policy for help: When you get big, you have to cut a deal w/big studios/music companies or go out of business because they will litigate you to death. Also: Consider Joe Singer’s interesting argument that in property, standards can be more predictable than highly complex rules. One reason his argument is relevant is that the incentive to litigate is huge in this context—the content cos behave as if it’s bet the company litigation and the ISPs know it is—and thus rules may be no better because compliance with rules is so often debatable.
Bridy: rules around notice and takedown isn’t unworkably complex—rules can become unworkable. But there’s less litigation about what a compliant notice is, but she thinks the law has become clear (not sure I agree) whereas “obvious to a reasonable person” is going to be a jury question.
Ginsburg: you could have more notice—in Europe, a closed system. Fair use is more flexible which is why some in this room advocate for it. Desire for more certainty = careful what you wish for.
Liu: certainly there is the persuasive argument that clarity and open texture of fair use may trade off. But he’d like to raise the freedom of movement.