 The cases we have considered thus far involve defendants who operate sites, dance halls, flea markets, and so forth, in which third parties engage in copyright infringement. Somewhat different and more difficult are cases in which the defendant sells a product or provides a service that its customers sometimes employ in ways that violate the copyright laws and sometimes employ in ways that do not. As I mentioned earlier, such cases have come to be known as, quote, dual-use technology cases. There have been four major cases of this sort in the United States. The Sony or Betamax case, which I discussed in connection with the Fair Use Doctrine in lecture number nine, a set of rulings concerning the Napster file sharing service, a decision by the Court of Appeals for the Seventh Circuit involving one of the successors to Napster, known as Imster, and a second decision by the Supreme Court in the Groxter case. I'll briefly describe the facts and holdings of those decisions. In the Sony case, as you already know, the defendant prevailed. In all three of the file sharing cases, by contrast, the defendants lost. However, the ways in which they lost varied considerably. Put differently, the courts adjusted and then readjusted the law of secondary liability to address the changing structure of the file sharing systems. A charitable interpretation of the sequence of cases would depict it as an illustration of the flexibility and wisdom of common law adjudication. An uncharitable interpretation would see it as a manifestation of what is sometimes called result orientation. In other words, a determination by the courts to suppress a social practice they saw as pathological and a willingness to contort the law to achieve that end. Which is the more apt interpretation I leave to you. The basic facts of the Betamax case, as you'll recall, are that the film studios licensed television networks to broadcast their films to viewers with the expectation that the viewers would watch the embedded ads and then buy some of the products promoted by those ads. The advertisers in turn would pay the TV network's fees, some of which the networks paid to the studios in the form of license fees. This long-standing business model was disrupted by Sony when it began manufacturing and selling VCRs, which among other things enabled the viewers to avoid watching the embedded advertisements, thus endangering the flows of revenue through the system. Because the viewers were using their machines to make verbatim copies of the studios films, the studios might have brought suit against the viewers, but the impractic ability of that approach prompted the studios instead to bring suit against Sony, the manufacturer of the devices the viewers were employing. In the end, as you know, the studios lost in the Supreme Court barely. The two interlocking rulings that enabled Sony to escape liability are set forth on your screen. The second ruling, the time-shifting constitutes a fair use and is thus a lawful use of a VCR we discussed in lecture number nine. Our concern here is with the first ruling. The manufacturer of a device that can be used to violate the copyright laws is liable for contributory infringement if and only if the device is not capable of substantial non-infringing uses. The key phrase comes at the end, quote, capable of substantial non-infringing uses, close quote. Using the first letters of those words as an abbreviation, this dimension of the Sony decision might be called the Casnu defense. This standard for secondary liability, which the court adapted from a loosely analogous aspect of patent law, is quite generous to defendants. Taken literally, it means that the manufacturer and distributor of a dual use technology does not need to show that his product is often used for legal purposes or even that it is ever used for legal purposes, but merely that it is capable of being used for substantial legal purposes. Sony, of course, easily passed that test because as we've seen, time-shifting, which the Supreme Court for the reasons we've already considered determined to be lawful, was the most common use of VCRs. But the way in which the court phrased the Casnu defense seemed to offer an escape hatch to defendants whose products were much less often employed legally. The relationship between the Casnu defense, announced and applied in the Betamax case, and the overall doctrine of secondary liability is not entirely clear, but it appears that the way in which the defense engages with the standard requirements is by negating the element of knowledge, which, as you know, is one of the requirements of contributory infringement. More specifically, the fact that a device is capable of substantial non-infringing uses prevents the plaintiff, in the court's words, from showing that the defendant, quote, has sold equipment with constructive knowledge of the fact that his customers may use that equipment to make unauthorized copies of copyrighted material, close quote. Now, in theory, the studios could have overcome that impediment if they could have shown that Sony had actual knowledge that a particular customer was using her machine illegally. For example, that a particular customer was using her machine for a librarian, which even the majority of the Supreme Court seemed to assume did not qualify as a fair use. The studios did not make such a showing and probably could not have done so. But this possibility, as we will see, played a role in some later cases. Another ambiguity in the Betamax opinion concerns whether the Cosnu defense applies only to claims of contributory infringement or also to claims of vicarious infringement. Almost all the court's opinion focuses on contributory infringement, but in one cryptic footnote, specifically footnote number 17, the court observed that, quote, the lines between contributory infringement and vicarious liability are not clearly drawn, and that reasoned analysis of respondents' unprecedented contributory infringement claim necessarily entails consideration of arguments and case law, which might also be forwarded under other labels, close quote. The court thereby left open the possibility that its ruling applied to vicarious liability as well. Again, this possibility didn't much matter in the Betamax case itself because in any event the studios could not show that Sony had the right and ability to supervise the ways in which its customers used its machines and so could not establish vicarious liability even if the Cosnu defense were inapplicable to such claims. But it would make a difference in some subsequent cases. So that's where things stood as of 1984. For the next 15 years, the Cosnu defense, announced in the Betamax case, stood as a powerful bulwark for firms developing products and services that were sometimes used to engage in copyright infringement. What destabilized this doctrinal structure was the rise of file sharing. The next three cases in the sequence struggled to reconcile file sharing with the Betamax framework. The first and arguably most famous of the three file sharing cases was Napster. I've mentioned the Napster system and the associated legal controversy before. Here's a somewhat more detailed description of the facts. Napster consisted of a website and an application program that together enabled music fans to exchange copies of sound recordings. The Napster website, which was www.napster.com contained a directory listing all of the recordings that could be found on the hard drives of all of its members' computers and an index to those recordings. A music fan who wished to join the community and gain access to those recordings would first log into the website and download from it a free piece of software known as Music Share. After installing the software on his computer, the user would sign up by selecting a unique username, typically fictitious, and a password. He would then create on his computer a user library, close quote, into which he would copy any recordings, typically in the form of MP3 files that he wanted to share with other users. He would then log into the Napster system. His software would talk to the software on Napster's servers. One result of that conversation is that a list of the files on his computer would appear in the Napster directory. The files themselves would not be copied onto the servers of Napster's site, just their names. That's the significance of the hollow circles in this diagram. Other users would be doing the same thing. Suppose one of the subscribers, say user number two, wanted to find some recordings by Eric Clapton. She would submit a search request using Clapton's name. If any of Clapton's recordings appeared in the index, the software would identify which of the libraries of currently logged in subscribers contained those recordings and provide that information to the application on user number two's computer. That information would enable her computer to connect directly to one of those host computers here belonging to user number one, download a copy of the file directly from the host computer and save it on user number two's hard drive. The fact that the two user's computers were in direct contact explains the name peer-to-peer copying. User number two could then either play the file directly from her computer or, if she had a CD burner and the appropriate software, she could convert the mp3 file to a wave file and copy it onto a CD, which could then be played on any CD player. Now if only a few people engaged in this practice, the owners of the copyrights in the compositions and recordings most likely would have tolerated it, just as they had tolerated the long-standing practice of teenagers making so-called mixed tapes for their friends. But the system grew extraordinarily fast. By October of 2000, roughly a year after its launch, it had 32 million subscribers. Four months later, it had 80 million. Nor was it limited to the United States. Indeed, the percentage of people with internet access, who used Napster, was higher in Canada, Argentina, Spain and Brazil than it was in the United States. Copyright owners, not surprisingly, became increasingly concerned. Soon after the launch of the service, the Napster executives contacted the record companies and sought to obtain licenses to distribute their works. The record companies considered such an arrangement. Indeed, Bertelsmann, the German parent company of one of the record companies, extended a loan of $80 million to Napster and tried to persuade the other record companies to license their catalogs to Napster. But in the end, the record companies decided to litigate rather than negotiate. As in the Betamax case, the companies could have brought suit against Napster's subscribers, who were actually engaged in copying their works and distributing unauthorized copies. But again, the impracticability of that option prompted the copyright owners to pursue Napster instead. Because Napster was not itself copying recordings or distributing copies, the owners did not have a good claim for direct infringement. But they could and did contend that Napster was secondarily liable for the conduct of its subscribers. Judge Patel of the district court found in favor of the owners. Napster appealed to the Ninth Circuit, which ruled in brief that use of the Napster system to sample songs is not a fair use. Because Napster is capable of a substantial non-infringing use, the owners of the system lack the constructive knowledge of infringing use as necessary to support contributory liability. Note here the impact of the Betamax ruling. So far, things look good for Napster, but its luck would not hold. Next, the Court of Appeals ruled that because the Napster operators can ascertain whether a particular recording is being copied illegally, they have timely actual knowledge of specific acts of infringement necessary to support contributory liability. Remember that this was an option left open by the Betamax case. The Ninth Circuit in Napster ceased it. Finally, for the same reason, the Court of Appeals ruled that the Napster operators have the, quote, supervisory control, close quote, over the conduct of its subscribers, necessary to support vicarious infringement. Death came swiftly. By the summer of 2002, the Napster system had been closed and the company filed for bankruptcy. Critics of the record companies argued that by suing Napster they made a catastrophic strategic mistake. Had they agreed to license the Napster service, they might have been able to collaborate in the construction of a lawful music subscription system, from which they could have earned a great deal, and avoided alienating millions of consumers. By adopting instead a confrontational posture, the record companies delayed for roughly a decade the emergence of such services. By then, the recording industry had shrunk by half. In the US, its gross revenues in 1999 were around $13 billion. Today, they're around $7 billion. Global revenues have shrunk at a corresponding pace. Today, they're around $16 billion. This collapse could have been avoided if the companies had embraced peer-to-peer technology instead of trying to suppress it. The record companies, of course, disagree with this assessment. Capitulation to Napster, they argue, would have led to an even faster decline in the industry, as other unauthorized and thus free services sprang up and drew customers away. Perhaps. It's unlikely we'll ever be able to resolve this controversy. So, back to our story. Napster's demise did indeed open the way for follow-on similar services. One of them was Amster, so named because it piggybacked on AOL's instant messaging service. It was similar to Napster in many ways, but differed in one. Encryption prevented the operators of the Amster system from knowing the contents of the files its subscribers shared. Nevertheless, the intentions of the Amster creators were reasonably clear from the fact that they provided users online tutorials, showing them how to use the system to exchange sound recordings, most of which would likely be subject to copyright protection. Like its ancestors, Amster attracted subscribers with extraordinary speed. It was launched on August 8, 2000. Within three days, 20,000 people had signed up. By early 2001, it had 2.5 million registered users. And by the end of April 2001, the number had grown to 4.2 million. The purpose of the encryption included in Amster's design was, of course, to avoid the kind of, quote, actual knowledge, close quote, of infringing behavior that had doomed Napster. And the judgment of the Court of Appeals for the Seventh Circuit, this was too clever by half. In the Court's judgment, Amster was plainly liable for contributory infringement. The Court brushed aside the cause-nude defense that had saved Sony on the ground that the defendants had failed to present evidence of any lawful uses of their system. The Court next ruled that the defendant's strategy of seeing no evil could not save them. In its words, willful blindness is the equivalent of actual knowledge. Finally, material contribution was established by the defendant's active encouragement of infringing uses of the system. Interestingly, the Court was less sure that the defendants were liable for vicarious infringement, but no matter, contributory infringement was sufficient. The holdings in the Amster case were relatively unsurprising. Much more eye-opening were some statements made by the Court in dictum suggesting doubts concerning the pillars on which the Betamax ruling had rested. Judge Posner, writing for the Court, expressed his disagreement with the Betamax formulation on two fronts. First, he contended that to trigger the cause-nude defense, a defendant must demonstrate that its product has substantial non-infringing uses, not merely that it is capable of substantial non-infringing uses. Adoption of the suggestion would entail converting the doctrine from a cause-nude defense to a cause-nude defense. Second, he argued that even if the defendant makes such a showing, if the infringing uses are substantial, the defendant must also show that it would have been, quote, disproportionately costly, close quote, to design its product so as to eliminate or reduce the infringing uses. These principles were not necessary to the Amster opinion, but they voted ill for the developers of dual-use technologies. The last of the major file-sharing cases involved the Groxter service. It differed from its predecessors, somewhat in structure. Specifically, the fast-track technology that it, along with some other companies, employed was decentralized. Not quite as decentralized as Nutella, a popular system developed by a renegade programmer at AOL, but more so than either Napster or Amster. In brief, a fast-track user interested in exchanging vials would locate, with the aid of a central server, one of a set of computers connected to the internet that functioned as so-called supernodes, coordinating search requests among clusters of users. Once engaged in the system, the user could submit a request, say, for a particular Clapton recording. If a copy of the requested recording were located on one of the connected user's computers, it would be delivered to the requesting party with no further involvement by a Groxter. Partly because of this structure, it was reasonably often for non-enforging purposes, such as distributing movie trailers, sharing the works of Shakespeare, and locating computer software for which distribution is permitted, as well as for exchanging recordings unlawfully. These circumstances prompted the district court judge who first heard the copyright owner sued against Groxter to deny liability. The Court of Appeals for the Ninth Circuit also ruled in favor of Groxter, reasoning that the existence of legitimate uses of the system means that the defendants could not be charged with constructive knowledge of unlawful uses, just as Sony could not be charged with constructive knowledge of unlawful uses of its VCRs. Groxter did not learn of specific illegal uses of its technology until it was too late to stop them. And unlike Fano visa, the defendants were not supplying the site and facilities for illegality and had no affirmative duty to alter their software to prevent illegality. For all these reasons, in the judgment of the Ninth Circuit, defendants were not liable for contributory infringement. With respect to Vicarious infringement, the plaintiffs had failed to establish, said the Ninth Circuit, the necessary right and ability of the defendants to supervise the direct infringers. These arguments, as I hope you see, are reasonably straightforward applications of the law as it existed at the time. The Ninth Circuit was not tilting the table in favor of the defendants. But the copyright owners and their supporters contended that if this is where the current law leads, the law must be wrong and should be modified. They asked the Supreme Court to revisit the Betamax case and to adjust the pertinent standard. How exactly? Various possible reforms were proposed. The copyright owners themselves contended that the Cosnu defense, announced in Sony, should not apply when, quote, the primary or principal use of a product or service is infringing. Close quote. The Solicitor General offered a different approach. Secondary liability should be imposed, quote, if the defendant's product is overwhelmingly used for infringing purposes and the viability of the defendant's business depends on the revenue and consumer interest generated by such infringement. One group of economists argued instead that secondary liability should be imposed if the defendant, quote, could eliminate or greatly reduce the level of infringement without significantly cutting down on the quantity and quality of lawful uses. You should hear, hear an echo of Judge Posner's proposal in Amster. The most radical approach was offered by another group of economists. Lower courts should consider all economic variables in particular cases when deciding whether to impose secondary liability. Other parties argued instead that the Betamax Cosnu defense should be preserved because, first, tighter standards for contributory vicarious infringement would endanger some socially valuable dual-use technologies. The unpredictability of all of the proposed alternatives to Cosnu, when combined with a threat of statutory damages a topic we'll examine next week, would chill innovation and the entertainment industry could be saved, they argued, through less radical forms of therapy. In the end, the Supreme Court split the difference. It left the Sony Cosnu defense intact. But it added to the set of forms of secondary liability a new doctrine, adapted from other sources, which it referred to as, quote, inducement. The key sentence in the Supreme Court's opinion announcing this rule is set forth on your screen. One who distributes a device with the object of promoting its use to infringe copyright shown by clear expression or other affirmative steps taken to foster infringement is liable for the resulting acts of infringement by third parties. The question immediately arises, how does a plaintiff go about proving inducement of this sort? The courts answer by offering evidence of, quote, purposeful, culpable expression and conduct. What kinds of evidence will suffice? Not that the defendant knew that his product or service was sometimes used for unlawful purposes. That was not enough to give rise to liability in the Betamax case, as we saw, and it's still not enough. Nor is it sufficient for a plaintiff to offer evidence that the defendant is providing customers generic product support. Something more is essential in order to demonstrate inducement. In its opinion, the court identified four sorts of relevant evidence. Advertising illegal uses, targeting customers known to engage in illegal uses, failure to adopt infringement-reducing technologies, and commercial sense of the enterprise that depends on illegal uses. The last two types of evidence, though relevant, will not be enough on their own, to succeed, it seems, a plaintiff needs some evidence of the first two types. Advertising unlawful uses or targeting kinds of consumers known to be engaging in infringement, such as reaching out to former Napster subscribers who are now looking for a substitute file sharing service. Like most Supreme Court opinions, the Grokster ruling contains some ambiguities. For instance, in one crucial sentence, the court indicated that liability under the inducement theory arises when the distributor intended and encouraged the product to be used to infringe. The two verbs in that sentence are different. The former suggests a subjective standard under which the crucial variable is the defendant's state of mind. Evidence pertaining to the defendant's intention would thus be relevant. The latter, by contrast, suggests an objective standard and only the defendant's conduct would be germane. These and other ambiguities are tacitly left to the lower courts to work out. But the principal implications of the Grokster ruling are clear enough. If you actively promote unlawful uses of your product or service, you're in trouble. Ironically, this standard, had it been in place in the late 1970s, might have led to a different result in the Betamax case. As you'll recall, I hope, Sony sponsored some advertisements that seemed to tout the use of its VCRs for librarying, not just time shifting. And librarying is an activity that the Supreme Court subsequently seemed to assume was unlawful. This appears to be the kind of active promotion of illegality that the court in Grokster condemns. Whatever the merits of the Grokster approach evaluating conduct in the past, it is likely to have diminishing force in the future. Why? Because attentive potential defendants will be sure not to advertise or otherwise promote the unlawful uses of their products and services and to avoid mentioning such things in their internal correspondence. In short, the inducement rule is likely to catch the unsophisticated creators of startups, not sophisticated actors with good lawyers. That's not an optimal legal standard. Thus far, we've been examining the expansion in recent years of the judge-made doctrines of secondary liability. But there's another force to be reckoned with. And economically and politically important group of businesses that are potentially vulnerable to liability under these expanding rules have sought and obtained qualified immunity. Specifically, in 1998, as part of the so-called Digital Millennium Copyright Act, commonly abbreviated as DMCA, a group of companies that provide a variety of internet-based services persuaded Congress to grant them protection against copyright liability. Both direct and secondary liability provided that they complied with some reasonably precise requirements. The complex rules giving these enterprises protection are now embodied in Section 512 of the Copyright Statute. The details of Section 512 are much too intricate to examine in this lecture. What I will do in the remainder of this segment of the lecture is to sketch the basic principles underlying this portion of the statute and to describe the most important of the judicial opinions that thus far has construed Section 512. A subset of the components of 512 are set forth on your screen. For the complete set, click on this link. Internet service providers, meaning passive intermediaries who simply carry packets of information from one party to another, are not liable for copyright infringement, provided that they do not select the stuff that is sent or who it is sent to or modify it en route. This should not be surprising. The postal service is also not liable if some of the letters it carries contain unlawful material. Things get less obvious when we shift from ISPs to so-called OSPs, online service providers, which are defined in Section 512 K1B as providers of online services or network access or the operators of facilities therefore. To qualify for the 512 Safe Harbor, such entities must abide by the requirements set forth in 512 I in brief adopting, announcing, and abiding by a policy for terminating people who repeatedly abuse their services and accommodating reasonable technological protection measures adopted by the copyright owners. If an OSP qualifies, it enjoys immunity with respect to various activities, including unsurprisingly passive caching of the material that turns out to be infringing. The most controversial of the activities for which OSPs may secure immunity concerns not carrying packets from one person to another or caching it for a faster retrieval, but rather storing material sent to them by one person and making that material accessible to lots of other people. In 1998, this practice was not terribly common, but it is now ubiquitous. OSPs who do this are free of liability if they first publicly identify an agent that copyright owners can contact to tell them that specific works are being stored on their systems unlawfully. Second, abide by detailed set of procedures for removing such material from their systems or resolving disagreements between the copyright owners and the posters concerning their legality. Third, lacks the kind of specific knowledge of illegality spelled out in section 512c1a. And fourth, lacks the kind of financial interest spelled out in section 512c1b. The clauses pertaining to knowledge and financial interest should sound roughly familiar to you. They pertain to the same kinds of concerns that have long figured in the doctrines of contributory infringement and vicarious infringement, respectively. But those concepts take here an unusual form and serve an unusual role. If an OSP accused of storing and distributing infringing stuff can show that he lacks both of these things, he's home free. If he cannot, he's not necessarily liable but must run the gauntlet of the doctrines of direct and secondary liability we've considered in the last four weeks. So much hinges on how exactly these requirements are construed. The case that to date has examined these requirements in most detail is Viacom versus YouTube. The basic facts are likely familiar to most of you. YouTube was founded in 2005 by three former employees of PayPal. The purpose of the initial version of YouTube was not entirely clear but its principal function soon became to enable people to post video clips and make them available for the world to watch. It grew extremely rapidly. By 2007, YouTube was the dominant site of its kind and its lead over its rivals continued to increase. YouTube's success was attributable in part to the sophistication and convenience of its technology and in part to the support it received from Google which purchased the company in November of 2006 in a stock for stock transaction worth roughly one and a half billion dollars. Last but not least, YouTube flourished in part because it was willing, particularly in the early years, to host videos consisting of or containing commercial copyrighted material which lots of users wanted to see. The owners of the copyrights in those materials reacted in various ways to the rise of YouTube. Some shrugged, some negotiated licenses with YouTube or Google permitting them to host their content and finally a few brought lawsuits. Included in this last group was Viacom, the entertainment giant that owned the copyrights in many television shows, episodes of which were uploaded to YouTube without permission by fans and then watched by other fans. The numbers were large. In 2007, when it initiated the lawsuit, Viacom identified 63,497 clips then hosted on YouTube that Viacom contained its copyrighted material. YouTube acknowledged that some of the videos on its site contained copyrighted material and promptly removed the 63,000 clips identified by Viacom. But having done so, YouTube asserted that it bore no further legal responsibility. Having complied with the so-called notice and takedown procedures that, as we've seen, are mandated by Section 512, YouTube contended that it was entitled to the benefit of the safe harbor of Section 512c and thus could not be liable for direct or secondary copyright infringement. Viacom, as you might expect, disagreed. Here's a statement released by Mike Fricklis, Viacom's general counsel, summarizing the company's stance. This paragraph on your screen succinctly states the arguments that Viacom hoped to use to deny YouTube the benefit of Section 512c. As you'll recall, that provision does not apply when a defendant either has actual knowledge of infringing material on its site or receives a financial benefit directly attributable to infringement which the defendant has the right and ability to control. Fricklis, as you can see, contends that YouTube fails to satisfy both of those requirements, either of which is fatal to its 512c defense. This paragraph, aimed more at a general public audience, explains why, in Viacom's opinion, it's fair to require sites like YouTube to purge their systems of infringing material rather than to require copyright owners constantly to monitor such sites. So those are the arguments. The district court sided with YouTube on all contested issues and granted its summary judgment motion. In 2012, the Court of Appeals for the Second Circuit reversed. The interpretations of the key provisions of Section 512 adopted by the Court of Appeals were, as we will see, highly favorable to OSPs like YouTube, but not quite as favorable as the interpretations that had been adopted by the district court. Thus, the Court of Appeals remanded the case for reconsideration in light of its clarified rules. Here, then, are the key portions of the Court of Appeals' opinion. First, the Court ruled that the disqualification contained in 512c1a1 is triggered only by actual, quote, subjective, close quote, knowledge of specific infringing material in the defendant's site, or willful blindness of the sort that doomed Napster. The disqualification contained in 512c1a2, the so-called red flag provision, is triggered only by actual, subjective knowledge of facts that would have made infringement objectively obvious to a reasonable person. Merely being aware that there's lots of infringing material on one site is not enough to trigger either of those provisions. However, the Court ruled Viacom had pointed to a few instances in which internal correspondence by YouTube executives suggested that they were aware of specific infringing files on their site, from which a jury might infer the requisite level of knowledge. Next, the Court of Appeals ruled that the disqualification contained in 512c1b is triggered only by proof of somewhat greater control over the infringing behavior than is required for ordinary vicarious infringement. Otherwise, the Court pointed out, the safe harbor of 512c wouldn't be of much value because it would give defendant's immunity only from conduct that would not trigger vicarious liability anyway. So what sort of additional control would cause a forfeiture of the safe harbor protection? Not merely the ability to block access to material posted on its service, but perhaps the sort of, quote, purposeful culpable expression and conduct, close quote, that, as we saw at the Supreme Court in Groxter, suggested would give lies to liability for inducement. Uncertain, on this score, the Court of Appeals suggested that the district court think about it further. Finally, the Court of Appeals construed 512c itself broadly, interpreting it to apply to all software functions performed by sites like YouTube, quote, for the purpose of facilitating access to user-stored material, close quote. The state of the law governing OSPs after the Viacom decision might be depicted graphically as follows. As you now know, OSPs are potentially liable, not just for direct copyright infringement, but also for any of the forms of secondary infringement, contributory, vicarious, and post-Groxter inducement. Section 512c not only gives greater specificity to the limits of those doctrine, but also, more importantly, allows OSPs to engage in some kinds of conduct that, in the absence of 512, would give rise to liability. In particular, hosting material without enough specific or red flag knowledge that particular pieces are infringing to trigger the disqualifications of 512c1a and without the tight level of control necessary to trigger the disqualification of 512c1b. Exactly what kinds of conduct fall into the zones identified by the arrows in this diagram will have to be worked out in subsequent litigation, but this seems to be the structure that OSPs must and may now rely upon, at least if the second circuit approach holds. This concludes our discussion of secondary liability and the especially troublesome issues presented by so-called dual-use technologies. After the break, we'll turn to a very different topic, technological protection measures.