Fourth Breakout Session
Copyright and Trademark Enforcement
Do ‘Groundless Threats’ Statutes Curtail IP Over-Enforcement?
Research interest: how disputes are negotiated and resolved outside of court, as most legal disputes are. Studying it qualitatively, based on interviews w/lawyers and sometimes their clients. Mapping the disputing process: what disputes arise, how does potential turn into actual dispute, how does it get resolved or enter the formal legal system.
One issue: are there IP bullies? Earlier work: enforcing lawyers’ own accounts indicate potential overenforcement because it works—when it makes business sense, they make very weak claims, and lawyers think it’s ethical and appropriate.
Some countries have a statutory scheme designed to deter inappropriate prelitigation IP threats, and so he decided to look at the disputing process in two jurisdictions with such schemes—Australia and England, w/laws dating back to the 19th century steam patent wars. (Now I want a steampunk fantasy based on the 19th century steam patent wars.)
Basic scheme: to deter bully-boy tactics. English version: Injunctions, declaration that threat is unjustified, or even monetary damages are available: even the lawyer who writes the threat can be sanctioned. Early findings: skeptical that statutes are effective.
It’s very difficult to police that area of social and legal life. It’s a problem outside of the formal legal system. The remedy is to enter the formal legal system, and there are obvious costs and risks to invoking that system. It may make more sense just to lump it, especially when you’re dealing with sophisticated parties with experienced legal counsel. The statutes are complicated; savvy lawyers can figure out tactical ways to use their ambiguities and technicalities so they can threaten without recourse. It is very rare to have published cases under these statutes, despite existing for over 100 years.
Critiques and amendments to these statutes have been made based on claimed unintended consequences. Alleged: encourages litigation, because one of the defenses under groundless threats is that you have already sued; absolute immunity thereafter, even if you don’t serve the lawsuit/go further. Lawyers have other tactics to get around the statutes. Compare: discovery sanctions in the US are virtually unavailable. Lawyers have incentives to tell the clients that the laws are highly complicated—but that just means that the clients need specialized IP counsel to get around them! So it creates a need for lawyers.
Goldman: are there empirical studies in Europe?
A: Not so much; the most recent English report went through history, perceived problems—wasn’t based on empiricism, but on lawyers’ opinions.
Leah Chan Grinvald: the law commission published a report in 2014. Some lawyers in that did say that the law impeded certain kinds of enforcement.
A: right, maybe the exceptions have missed the problems. The lawyers say two different things: they say to clients that the results clients want are available if they’re careful.
Matthew Sag: declaratory injunction standard? Role of fee shifting may also be relevant.
A: needs to find that out in further research.
Ramsey: categorical statutory rules (comparative advertising etc.)—do they matter to what claims do or don’t get brought at all?
A: also more research—need to be careful about the statutory context.
Tyler Ochoa: fee-shifting can also deter use of groundless threats statutes too.
A: that’s why he’s interested in one-timers versus repeat-players. One-timers seem less willing to take risks. Repeat targets have sense of when they should use the legal system and when not; one-timers worry about risks of fee-shifting.
The Criminal Copyright Gap
Criminalization increase over time. More recently: analog high tech phase, then digital high tech phase. Vast amount of legislation increasing scope and punishment, incl. 1997 NET Act arguably criminalizing end user infringement. More types of works, more types of actions, harsher punishments.
Is there a paradigm shift from civil orientation to criminal orientation? Looking to enforcement, there are numerous units now designed to handle enforcement, e.g., Office of US IP Enforcement coordinator. But that has not resulted in a consistent actual increase in criminal cases. Since 1986 there’s a big rise in criminal filings, but copyright cases haven’t taken a larger share of those filings. More defendants, but still an enforcement gap.
Why the gap? (1) Int’l agreements provide for enforcement, but can’t specify the level thereof. Obliged to criminalize, but not obliged to enforce. (2) Prosecutors aren’t particularly vulnerable to pressure from this group v. others. (3) Deterrence—maybe the prospect of penalties alone worked, and enforcement isn’t necessary. Not sure how plausible this was, but House Report said that NET Act was just supposed to enable a few more prosecutions per year. (4) Tolerance rate? Rightsholders willing to handle their infringements through the civil system in most cases. (5) Enforcement problems. Trying to prosecute someone for $1000 over 6 months requires analyzing a database—when one household shares 1 IP address, you can’t tell who infringed. Prosecutors also don’t know the total retail value. P2P networks are slowing down because people have various methods of infringing that are basically now very hard to detect. Criminalization is mostly from P2P, and if you’re not using that it’s very hard to prosecute. (6) Gov’t guidelines: until 1985, there were no guidelines for prosecuting infringement, but that was supposed to change/get same attention as other felonies, though it didn’t. (7) Prosecutors might not view it as important v. other offenses. Resources.
A: mainly targets bootleggers/CD sales on street; we don’t have the NET Act. Only larger operations like streaming websites are targeted.
Grinvald: on being hard to track: what about © trolls with mass filings against swarm, then use subpoenas to get more info.
A: haven’t seen that in criminal copyright.
Q: not surprised in gap—institutional environment drives enforcement rate. Variables in cross country analysis might help.
A: trying to examine US because each country is unique. Gap surprised him given congressional/legislative focus on getting more and higher fines. Only on the books, though.
Q: take a look at revolving door phenomenon b/t agencies and copyright owner industries. Get attention of prosecutors in ways that other criminal victims can’t—Megaupload prosecution, supported/instigated by copyright owners. (Also Dajaz seizure?) Given the amount of regulatory capture why isn’t there more of this?
A: interesting question for him too. Hearing: Disney was asked if it had sought prosecution, and counsel said yes, and they’d been turned down based on limited resources. But that was 14 years ago. Political compromise, given interests on other side: law on the books enough?
Ochoa: consider overall level of prosecution. We have a very large prison population (though mostly state)—so many more prosecutions available that resources are put elsewhere—terrorism.
A: thinks that’s main issue.
RT: consider structural racism. It’s not terrorism cases that are occupying federal prosecutors’ time. It’s drugs. © infringement is stereotypically white and nerdy, while there’s more support for laws that disproportionately affect blacks.) Compare drug use/other “victimless” or non-physical crime.)
Criminal Trademark Enforcement and the Problem of Inevitable Creep
Focus on TM Counterfeiting Act, federal version, though there are state level laws and issues. There are many reasons to worry about state versions, but focused on federal b/c it’s an often ignored area, e.g. many people (including me!) skip it in the TM course.
Supposed to target most egregious forms of infringement—use of identical mark for identical goods listed in a federal registration, where core confusion is at issue. Civil remedies are created to mirror this—some of what we’re seeing here may be spillover of civil rules into criminal side. Claim of supporters: criminal sanctions were necessary because civil sanctions wouldn’t deter counterfeiters who regarded liability as cost of doing business, and increased damages weren’t enough because of the health and safety risks of counterfeiting (now funding terrorism has been added to the list of dangers).
Limited, narrowly tailored was the claim: state of mind requirement—must intentionally traffic in goods/services and knowingly use a counterfeit mark to trigger the statute. Definition of what counts as counterfeit mark: substantially increased level of similarity was supposed to be required. Has to be “identical with or substantially indistinguishable from” a mark registered on the Principal Register. Civil context doesn’t require registration or high threshold of similarity. By contrast infringement liability is designed to work in cases where the similarity isn’t that great.
Statute also required likely confusion. But these requirements have all been relaxed along every dimension you can imagine. For example, post-sale confusion is accepted as relevant confusion for purposes of liability. Knockoff luxury goods scenario—maybe that doesn’t surprise you. Also some customs cases Congress was considering as part of the context involved knockoff luxury goods. To the extent that the justification for criminal remedies is health/safety, post-sale confusion really doesn’t fit the mold. Particularly pernicious effects at sentencing, because gov’t often argues that restitution should be based on full retail value of genuine goods, and that there’s a 1-1 substitution between genuine and counterfeit.
Courts have also watered down the substantially indistinguishable standard. Congress made a bunch of references to a particular case involving Rolex. That case said: standard should be that the average buyer would, if familiar w/P’s mark, conclude that the infringing mark was in fact P’s mark. Should be case by case, but the case law previously had been restrictive: Amazona v. Amazon not enough, and P superimposed over V wasn’t enough when the mark was LV’s L superimposed over V. The cases now conflate likely confusion with substantial indistinguishability—likely confusion becomes the test, collapsing into infringement test. Cases also mush the two the other way: if there’s substantial indistinguishability, it then follows that the likely confusion test has been satisfied.
Courts compare the wrong thing: you’re supposed to compare the allegedly counterfeit mark with a registered mark, but sometimes they compare a composite mark to two separate registrations and find similarity. Burberry case: plaid mark plus figural mark. Court used market availability to determine that it was counterfeit, instead of looking solely at the registrations as required by statute.
Courts also compare the wrong way: assess similarity question from perspective of use in the market, rather than objective side by side comparison, then do likely confusion in the opposite way by looking side by side instead of market context. Both of those are exactly backwards. Objective similarity is supposed to be distinct from market context, but likely confusion isn’t.
We are moving towards full overlap with infringement as courts relax standards. Whether you’re concerned about overenforcement of counterfeits or not, you should be concerned because of the greater IP context (even the greater federal criminal law context). Statutes enacted on the promise of narrow tailoring never hold narrow and cover lots more ground. CFAA; civil seizure of domain names; Economic Espionage Act.
Most radical claim: such lawmaking shouldn’t happen. If benefits exceed costs only if the law is narrowly targeted, you can expect that narrow targeting won’t work and your cost-benefit analysis should recognize this.
Less radical: if you do this (for example, revenge porn), you need to think very hard about how you’re going to cabin it, and we should search for limitation rules that work.
Grinvald: are there opinions smacking down the prosecutors?
A: only one routine pushback: when gov’t tries to prosecute use of TM on refurbished goods as counterfeiting. That’s the only one—the other examples are appellate courts broadening the law.
Ramsey: What about more categorical limits/exclusions.
A: possible. He worries because Congress had traditional uses of TM in mind, but it’s frequently used for product design. That’s a short step from enforcement against bag w/parts of design to use of product configuration—Indiana case said toy gun was counterfeit of real gun. Courts in criminal contexts have no concept of functionality, etc. Some categorical defenses might help. Different but related problem in civil context: Almost every TM P pleads civil counterfeiting in run of the mill TM cases, and the concept of counterfeiting has lost meaning as a separate category.
Jim Gibson: CFAA is a good example. But what if the harm is greater in forfeiture, seizure which we never see in the case law. Certainly at the state level, grey market stuff is taken off the streets regular.
A: domain name cases—some involved resellers, which he teaches his students is ok. Prototypical criminal TM infringer is a Chinese immigrant, almost always defended by PD who knows nothing about TM law. All happening at plea bargaining stage. Even more incumbent on courts to be rigid.
Ochoa: would be interested in followup: how often is this statute really being used? More often than ©, but maybe not all that often. That could catch plea bargains.
A: may be true. Wonders whether, when there aren’t a lot of prosecutions, what that says about how you should feel about breadth—maybe makes them more arbitrary.
IP Liability and Policy
Knowing How to Know: Secondary Liability for Speech
[missed intro] We talk about burdens on ISPs, but not about what it means to know. In corporate law there is discussion of what it means for a corporation to know through its employees.
One view might be: all you need knowledge of the activity that later turns out to be unlawful. Turns up in Goldstein and a few cases. Theory is that infringement is strict liability; why should it be different for contributory infringer?
Disagrees: should have some requirement of knowledge of infringement. Black dissenting in Sony: says that for injunctive purposes there’s always knowledge because infringement has been determined.
Law suggests there needs to be some awareness of character of speech for secondary liability to attach. But how? Courts have focused on means of acquiring knowledge. Notice often equals knowledge. At least when it’s DMCA compliant.
Her suggestion: talking about “knowledge” in important contests is a misnomer. Whether something is infringing/defamatory is a legal status, that is determined by a court that has the power/ability to declare that status and have it respected going forward. Something is not infringing or defamatory, as a fact, until that adjudication happens. When we say “did they know it was infringing?” we are asking “at the time of activity, what is the likelihood that a court would find that material infringing (or defamatory)? Did the D engage in a proper risk assessment of what a court was likely to do at some future date?” Saying something is fair use doesn’t make it so; any lawyer will couch advice in terms of prediction. Letter from © owner isn’t “knowledge” of infringement; they’re not always accurate in their assessments of what infringes.
We mean prediction, not knowledge. A theory of justified true belief can help. In order to say you “know” something, there has to be a tie between the facts you have and the likelihood of something happening in the future. We rely on predictions all the time: if the announcer at the train station says a train will be 20 minutes late, we generally believe that the train will be (at least) 20 minutes late, without considering the announcer’s reputation for truthfulness or the likelihood that an unauthorized person has taken over the PA system.
Thinking of it this way can remove us from a binary system and give us more of a scale—looking at the nature of the info available at the time to the person. May help mitigate hindsight bias.
A: That does make distinctions between knowing/having reason to know/should have known etc. Her project is to parse out those distinctions.
Tun-Jien Chiang: agrees w/basic epistemological claim about nature of knowledge, but predicate may not be necessary: jurisprudential claim about the nature of law may be unnecessary. When you say that there’s no murder until a judge determines that’s what my killing with malice aforethought was, you’re making a very strong claim. Do you really need that predictive model for your ultimate conclusion?
A: we’re not basing liability on the knowledge of having committed murder, we’re basing guilt on the fact of having committed murder. But here it’s different.
Followup: right, but you’re making a claim about what it means to know as well as a claim about what law is.
A: may not be necessary, but predicate of liability is knowledge of something that does not in fact exist yet.
RT: Same thing as Chiang (it’s not murder without the mens rea, and mens rea is what the court determines it to have been based on what you were thinking at the time—whether it’s malice aforethought is a legal conclusion; I think, under Heymann's model, that what you were thinking is the nonlegal facts that may or may not constitute malice aforethought). Suppose I believe, contrary to your intermediate claims, that there are “wrong” court decisions, and thus fair use or infringement is time-invariant. This has important political valence w/r/t fair use. Does your end argument still work? I think it can.
Thoughts on the Vimeo case, where the court found that awareness that a whole popular song was present counted as actual knowledge of infringement?
A: That’s an instance where the court is using a prediction model—the likelihood might be very high under those circumstances.
Site-blocking Orders in the EU: Justifications and Feasibility
[missed intro; darn these long hallways!] Hosting providers can do takedown; other entities may be able to block access—the Q is their responsibilities. Search engines: where should we consider them to fall? Search engine as data controller in recent EU privacy decision: obliged to remove certain results on the basis of a search on the person’s name.
DNS blocking: easy but easy to circumvent. IP address blocking: similar, higher risk of overblocking with shared IP addresses. URL blocking: most targeted but costly. Packet inspection (including direct packet inspection, DPI). Hybrid: IP blocking coupled with DPI as possible solution.
What criteria should be taken into account? Speed of implementation, cost, blocking effectiveness, difficulty of circumvention, compatibility w/ judicial process, integrity of network performance, impact on legitimate services. Site blocking injunctions are allowed to comply w/EC Directive on Copyright, which provides for injunctions against intermediaries. EC Directive on Electronic Commerce provides exceptions and limitations on enforcement. Directive on IP Rights Enforcement of 2004 is also relevant.
CJEU has been asked to interpret: service providers have no general obligation to monitor. Most recent CJEU case: UPC v. Constantin Film—whether an injunction should specify specific measures to be taken. Injunction issued mandating DNS and IP blocking for films; then higher court issued injunction without ordering specific technical measures. ISP appealed, and CJEU said that failure to specify is ok when the ISP can avoid incurring coercive penalties by showing it’s taken all necessary and reasonable measures, provided that internet users’ lawful access isn’t jeopardized and that the prevention of unauthorized access was made difficult/unlikely. (Gee, thanks.)
Various examples of different practices within member states. Some have found that site-blocking has proved ineffective. But injunction without specifics doesn’t seem compatible with common practices in most states. DNS blocking has been ordered by Spanish and Belgian courts, while hybrids have been ordered by UK courts.
Pros and cons of issuing injunction without specifics: It’s understandable that courts give intermediaries freedom to decide the means to be used given their technical knowledge—gives them flexibility. But ISPs may not be in a position to foresee the appropriate balance between copyright and users’ rights/right to conduct a business. Unbearable burden on service providers to determine specific measures and prove their suitability, specificity, appropriateness, and reasonableness. Result: uncertainty/inconsistency. Need for harmonization.
Suggestions: use formal scheme of injunctive relief based on relevant criteria, similar to what exists in DMCA; enhance principle of proportionality and strike a balance; provide general guidance on determination and deployment of proper technical measures; where necessary, use a court appointed expert to recommend specific technical measures on case by case basis; incorporate complementary measures into site blocking system to improve effectiveness/limit overblocking, for example by implementing notice and takedown w/in member states and delist results from search engines—establish coherent and systematic framework for reducing online copyright infringement.
RT: Not sure this description of “pros” is fair. The ISPs don’t seem to want this kind of freedom. Isn’t the “pro” that copyright owners get more of what they want, which at least they value? We can definitely call that a pro if we want, but I think we should be honest about what it is—not a benefit for the ISP at all.
A: true, ISPs don’t seem happy with having the burden of balancing.
Ramsey: note criticisms of US DMCA. Do you propose any remedy for people who are harmed by the blocking injunction?
A: ISP is supposed to ensure other users’ rights are protected. [I think this is part of the problem, though. The ISP isn’t a good balancer. Why isn’t there a remedy against the copyright owner who caused the problem?]
Q: why would specifying measures make them any more technically feasible? We are already seeing delisting orders—but incoherence nationally v. globally—can they be made consistent?