From Immunity To Regulation:
Turning Point Of Internet Intermediary Regulatory Agenda
After twenty years’ golden time for Internet intermediaries, who were offered immunity and “safe harbor” protection from liability for the illegal content originated with third-parties, recent cases saw a possible turning point towards more regulation imposed on them. Both in the U.S. and the EU, the courts started to be careful to provide broad protection for Internet intermediaries, who had become dominant platforms instead of fragile new means of communication they used to be. Before we adopt any reform, a comprehensive understanding of the rationales to support the existing legal frameworks and the reasons why they are no longer valid is critical and essential. In this article, I argue that free speech, industry development and Internet exceptionalism are the three main rationales supporting Internet intermediary immunity and “safe harbor” protection. However, as the Internet developed, the role of Internet intermediaries changed, which led to the gradual cessation of enjoyed privileges.
- The Problem Of Intermediary Liability
- Definition of Intermediary and Its Liability Problem in the Internet Era
- Development of Internet Intermediary Liability Legal framework
- The Rationales To Support Internet Intermediary Immunity
- Free Speech
- Innovation and the Industry Development
- Internet Exceptionalism and Self-governance
- Internet Intermediary Becomes the Potential Threat to Free Speech
- Innovation is Impeded: Collusion Among Oligarch
- Internet Intermediary Is Not the Exception
In the early days of the Internet, it was widely accepted that Internet intermediaries should be protected from sovereign power and left self-governed. However, the recent accusation against Facebook about its systemic political bias underscores a critical question concerning the growing impact of Internet intermediaries on society. As Internet service providers (“ISPs”) increasingly become the so-called chokepoint of online communication, to what extent should they be regulated?
In recent years, there has been a series of cases in which Internet intermediaries were held liable for content uploaded by third-parties, a trend that is surprising those who have long advocated intermediary immunity around the world. The Communications Decency Act (CDA) and the Digital Millennium Copyright Act (DMCA), enacted in 1996 and 1998 respectively , followed closely by the E.U.’s E-Commerce Directive (ECD), established a legal framework that offered broad immunity or “safe harbor” to Internet intermediaries from illegal third-party content. . However, recent cases indicate changes are making way. In Doe v. Internet Brands, Inc. and Delfi v. Estonia, both websites were refused the privilege based on CDA and ECD, respectively, as a defense to negligence claims. In Fair Housing v. Roommates.com LLC, the majority in the 9th Circuit aptly reflected this new direction:
The Internet is no longer a fragile new means of communication that could easily be smothered in the cradle by overzealous enforcement of laws and regulations applicable to brick-and-mortar business. Rather it has become a dominant means through which commerce is conducted. And its vast reach into the lives of millions is exactly why we must be careful not to exceed the scope of the immunity provided by Congress and thus give online businesses an unfair advantage over their real-world counterparts, which must comply with the laws of general applicability.
Given these new developments in Section 230 jurisprudence, several important questions naturally arise: After nearly twenty years since Section 230’s enactment, have ISPs found themselves at a turning point? That is, will judicial interpretation of Section 230 eventually shift from broad immunity to more regulation? If so, what are the causes behind such a transformation and why might increasing regulation be necessary?
Early critics and their accompanying propositions for reform of the current legal framework tended to focus on the congressional intent and the broad interpretation by the courts. Most commentators criticized the court’s conclusion in Zeran v. America Online, Inc., which they thought could not reflect congressional intent because it foreclosed adequate legal remedies for private individuals harmed by defamation or infringements. However, as the principles outlined in Zeran became widely accepted, the original critics shifted their focus on the greater implications of offering such broad immunity to Internet intermediaries, suggesting amendments intended to effectuate policies of efficiency and cost allocation. Thus, reforms such as “censorship by proxy” or “chokepoints of control” were proposed; these reforms involved calling ISPs into the “service of the law” by imposing vicarious liability.
Despite the evident merit and insight underlying these critics’ comments about Section 230 interpretation, most of the proffered arguments are problematically one-sided. That is, they were mainly based on analyses from given premises or goals to a specific solution, e.g.,efficiency of law enforcement or absolute right of free speech, which usually failed to balance competing values because of ignorance of the contexts of the legal framework. Those who criticized the inadequate legal remedies for the victims did not see the positive aspects to promote innovation. Likewise, those who advocated “censorship by proxy” reform overlooked the serious threat to free speech of such a reform. Put simply, most of the Section 230 reforms that have been offered so far are largely skewed against the compelling public interest that sits on the opposite side of the balance beam. Ergo, in this article, I will argue that a comprehensive and context-based analysis is essential to understanding the benefits and struggles of current legal framework for Internet intermediary immunity. By closely tracing the Internet industry’s rapid progress and its increasing impact on society, I will suggest that the Internet intermediary immunity legal framework is losing the factual basis that was once its supporting rationales. This factual and foundation shift suggests a future reform in the regulatory agenda.
This article proceeds as follows. Part I traces the history behind the development of intermediary liability regulation and the surrounding legal framework as well as introduces its main contents. Part II outlines and assesses the three primary rationales that support the current legal arrangements, which, at present, offer immunity or “safe harbor” for Internet intermediaries from tortious liability from unlawful third-party content. Part III articulates the three big changes, brought about by the rapid progress of Internet, that pose significant challenges to the former rationales after twenty years’ development.
I. The Problem Of Intermediary Liability
A. Definition of Intermediary and Its Liability Problem in the Internet Era
Generally speaking, an intermediary is “any entity that enables the communication of information from one party to another.” Based on this definition, any tangible artifact–a writing, a painting, or a sound recording–is by itself an intermediary that facilitates the communication of ideas or expression from one person to another. Newspapers, bookstores, libraries, as well as the recent digitalized information service providers including telephone companies, cable companies and Internet service providers are all intermediaries too.
Intermediaries function as a mediator between parties. However, as intermediaries become increasingly pervasive in daily life, a growing concern for policymakers is whether and to what extent these entities should be liable when others use their services to engage in unlawful conducts. Traditionally, we refer to common law to answer (at least in part) this question.
According to the defamation law, speech intermediaries are divided into three categories: publishers, distributors, and conduits. Based on their variant impacts, these speech intermediary are associated with tailored sets of liabilities. To illustrate, a party that, by analogy, most resembles a newspaper, which acts as a publisher that edits and controls the communication, is held legally responsible regardless of whether it was specifically aware of the materials at issue. Factors that indicate publisher liability include evidence of exercising editorial control and judgment over the choice of the materials published. At the other end of the spectrum, entities that are mere conduits of third-party content are not liable for the content they carry, even if they have, in a particular instance, actual knowledge of the content and its implications. One justification for such immunity is that the conduits, like a telephone company, can either not exercise editorial control over the contents they carry, or are restricted to refusing services to specific customers. Between publishers and conduits lie the distributors who exercise a certain control over the contents, such as refusing to distribute specific materials. With respect to distributors, which are most like bookstores and libraries, it is widely accepted that they are subject to liability “if, but only if, he knows or has reason to know of its defamatory character.” The rationale for imposing a lesser standard of liability for distributors is supported by Smith v. California, a case in which the court clearly stated the policy concerns over “the bookseller’s self-censorship” and “the impediment for the distribution of all books.”
Although the categorization of the differing intermediaries into the three strains (each with its own requirements for finding legal liability) appears clear-cut, it is somewhat difficult to apply practically. For example, actual or constructive knowledge of the illegal third-party content, while certainly a prerequisite for distributor liability, may not be sufficient. Additionally, if knowledge of falsity is required, then the liability of distributors is arguably hard to distinguish from that of conduits. Thus, the questions remain as to whether an intermediary should be considered a distributor or a conduit and what kind of liability should be imposed in this scenario. That being said, problem was not a critical one in the pre-Internet era, when publishers, like newspapers, controlled most mass-produced contents and thus took reasonable responsibility for those contents. However, when it comes to contents of the Internet era, the problem becomes a challenging one.
According to the Organization for Economic Cooperation and Development (“OECD”), Internet intermediaries can be classified into three categories: Internet service providers, search engines, and participative networked platforms. ISPs offer access to the internet, connecting different parties to form a given online communication. Search engines link related websites in connection with search inquiries, while participative networked platforms mainly host or cache contents uploaded by users. It is unclear whether an ISP can be considered as a conduit if it refuses to follow the network neutrality principle and discriminates against traffic and applications. It is also unclear whether a search engine can be deemed as a distributor if search results are biased by advertisements. As for participative networked platforms, it is still difficult to clarify their roles, especially as they induce users to upload contents. To summarize, the categorical boundaries among publishers, distributors and conduits are increasingly blurring as the number of Internet intermediaries rises, resulting in an inconsistent understanding among courts in the early cases, which we discuss in the following section.
B. Development of Internet Intermediary Liability Legal Framework
Confronted with the problem of liability, courts, in cases concerning Internet intermediaries, doctrinally diverged, as reflected in the Cubby, Inc. v. CompuServe, Inc. and Stratton Oakmont, Inc. v. Prodigy Services Co. cases.
The court in Cubby ruled that Compuserve was a “distributor” and could not be held liable for defamatory material if it had no knowledge of the illegal contents uploaded by its users. However in Stratton, the intermediary was ruled a “publisher” and thus was held liable, regardless of whether it had knowledge of the contents. Beyond Cubby and Stratton, which centered on defamatory contents, cases involving copyright infringements issues faced similar conflicts. Religious Technology Center v. Netcom and Playboy Enterprises v. Frena are two examples of such cases. In Netcom, the court ruled that Netcom could not be held liable for infringing materials posted by its clients. However in Frena, George Frena, the operator of a subscription computer bulletin board service, was held liable for the infringing photos uploaded by subscribers.
Although there are slight variations in different cases, the basic questions are the same: Should Internet intermediaries be held liable for unlawful contents that originate from third parties? If so, what kind of liability should be imposed? The conflicts between these cases have caused serious confusion about how to solve these problems. As the Internet developed, intermediaries such as search engines and social networks emerged and functioned in an even more complex way than outdated computer bulletin boards. These conflicts had to be settled in order to boost the burgeoning industry. However, it was not until the turn of the millennium that the legal framework offering Internet intermediaries limited liability or even immunity was established..In the United States, the Communications Decency Act (“CDA”) was enacted in 1996. Section 230(c)(1) of the CDA states that “No provider or user of an interactive computer service shall be treated as publisher or speaker of any information provided by another information content provider.”( The CDA does not extend its coverage to copyright infringements. Instead this was addressed by the enactment of the Digital Millennium Copyright Act (“DMCA”) in 1998. Specifically, Section 512 of the DMCA provides “conditional safe harbor from liability” as long as intermediaries do not have “actual knowledge” of infringement, do not directly benefit from the infringement, and do have a notice-and-takedown policy in order to be granted for the legal immunity. The EU legal frameworks followed the footsteps of the U.S. regimes. The Electronic Commerce Directive (“ECD”) adopted the basic idea of Section 512 of DMCA, offering safe harbors from liability for specific intermediary activities. However, the EU differs from the U.S. in its so-called horizontal approach, applying safe harbor to cover any kind of unlawful contents, including copyright infringements or defamations.
To reinforce the legal frame, the courts chose to interpret the CDA broadly. In Zeran, the 4th Circuit ruled that knowledge-based distributor liability was a subset of publisher liability and therefore was also foreclosed by Section 230.__ Although subsequent commentators pointed out some factors that made Zeran controversial, that controversy was not reflected in the trend of judicial decisions, which overwhelmingly followed Zeran’s holdings.
According to legal scholars, the broad immunity that the legal frameworks (i.e., the CDA and the DMCA) offered to Internet intermediaries was a “policy choice,” one which bypassed the chance to strike a balance between different values and instead merely chose to protect Internet intermediaries. Given the irreparable loss to the copyright holders from infringements and the irremediable harm to persons defamed, strong rationales supporting the intent of the congress and courts were needed.
II. The Rationales To Support Internet Intermediary Immunity
By offering immunity or safe harbor for unlawful contents originating in third parties, Internet intermediaries were given the privileges against Internet users who became victims in such cases. Given their vital roles in encouraging the development of Internet as a commercial and political resource, the current legal frameworks were championed as the “cornerstone of Internet freedom.” However, every coin has two sides. The protection for Internet intermediaries had has also proven to be ripe for exploitation; this is best seen in cases that involve revenge porn, child pornography, and defamation. While attempts at legislative reform have been minimal at best, and “legislative debates in the US seemed pretty much settled for around two decades now,” we can still argue confidently that “cyber-tort law is not settled until it is settled right.” Given the massive criticism and proposals for reform, it is critical to review the rationales supporting the current legal frameworks. Through understanding these rationales, we can clearly see congressional intent when the laws were enacted twenty years. The rationales can be understood in three categories: free speech, innovation and neutrality. However, as will be explained below, whether their validity remains today is in doubt.
A. Free Speech
Although it has been widely accepted that the rise of Internet promotes free speech, the positive impacts of the Internet cannot be gained without prerequisites. The immunity granted to Internet intermediaries is such a prerequisite because of the concern for “collateral censorship.”
Collateral censorship occurs when an intermediary suppresses the speech of others in order to avoid the imposition of liability on it due to that speech. Imposing liability on an intermediary could result not only in the suppression of unlawful speech, but also in overreach to the lawful contents, causing a “chilling effect.” Speakers can rarely capture the full social value of their speech, causing the externality problem. This is more serious for intermediaries than for original speakers. The latter, speak for all kinds of reasons, from monetary rewards to reputation aggregation, none of which accrue to the benefit of intermediaries. Thus, if the threat of liability casts on intermediaries, it is rational to expect them to steer clear from the unlawful zone as they have “a peculiarly fragile commitment to the speech that they facilitate.” Additionally, although intermediaries obtain benefits in other ways such as advertising with user-generated content, they are insensitive to the value of any specific piece of content, meaning that they lose little for deleting such content.
The collateral censorship problem applies to traditional intermediaries such as newspapers and book publishers, not only to Internet Intermediaries. However, immunity is only granted to Internet Intermediaries. The editorial control traditional intermediaries had over the contents partly explains the difference between the intermediaries. However, focusing on the communication environment might be more impactive. According to scholars, who hailed the Internet in its early days, Internet intermediaries dissolved the restrictions on speech and even provided “cheap speech” for the marginalized who would never have had the chance to be heard via newspapers. The rise of Internet intermediaries enabled individuals to speak directly to the masses without having to rely on traditional intermediaries, who had long determined the substance of media content. Thus, Internet intermediaries were claimed to have perfectly realized the promise of the First Amendment. The Supreme Court’s decision in Reno v. ACLU, echoed similar themes, stating that “through the use of chat rooms, any person with a phone line can become a town crier with a voice that resonates farther than it could from any soapbox. Through the use of web pages, mail exploders, and newsgroups, the same individual can become a pamphleteer.”
Given the dramatic shift, collateral censorship is much more severe for Internet intermediaries than their traditional counterparts. When Internet intermediaries are liable for online contents, they are likely to build new barriers and cripple ordinary people’s ability to speak online. In contrast with the traditional intermediaries, Internet intermediaries promote more freedom , which supports the rationale to offer privileges of immunity or safe harbor.
While collateral censorship is not the only rationale, it is considered to be the core factor undergird the limited liability legal framework for Internet intermediaries, both in theoretical literature, and in case law.
B. Innovation and the Industry Development
In Configuring the Networked Self, Julie Cohen argues that “gaps and inconsistencies within the system of legal rights, institutional arrangements and associated technical controls…protect the play of everyday practice[,]” which “create[s] opportunities for experimentation by a wide variety of participants where creative practice flourishes.” Following Cohen, Balkin comments that “immunities or safe harbor rules for intermediaries create discontinuities in digital enforcement regimes.” In their views, the current legal frameworks characterized by limited intermediary liability allow people to play with information and culture, thus fostering innovation in the gaps of the scope and coverage of copyright law.
Besides intentionally providing gaps where Internet intermediaries are free to experiment, the legal framework also reduces the innovation costs, which further promotes the generative nature of the Internet. By criticizing the gatekeeping theory, Zittrain clearly states his concern over the burdens imposed on innovation by holding intermediaries liable for third-party content. He argued that Internet intermediaries, like chat rooms or message boards who were incapable of coping with monitoring costs, could be induced either to “shut down entirely” or “to raise drastically the cost for their services.”
Similar ideas were also explicitly expressed by Anupam Chander. By focusing on the legal framework, Anupam proposed a new explanation for the rise of Silicon Valley.
U.S. authorities…acted with deliberation to encourage new Internet enterprises by both reducing the legal risks they faced and largely refraining from regulating the new risks they introduced.
It is not only the political concern for free speech and economic rationale for innovation that determine the current legal framework, but also a longstanding utopian culture propagated by cyber libertarians.
C. Internet Exceptionalism and Self-governance
The argument for the current legal framework provided for the Internet intermediaries has a close relationship with the long-lasting belief of Internet exceptionalism. Since the Internet began to reach the masses in the early 1990s, Internet exceptionalism emerged as a popular alternative for Internet governance proposed by so-called “cyber-libertarians”.According to these cyber-libertarians , the particular architecture of Internet distinguished cyberspace from the real world. This distinction has laid the foundation for Internet exceptionalism.
Because of a series of basic communication protocols, including TCP/IP and WWW, the architecture of Internet was characterized by open structure and non-discriminatory data transfer. As David Isenberg stated, the network was simple, or “stupid”, whose fundamental feature was neutrality among data packets, resulting in the inability to discriminate lawful data from unlawful data transferred on the network. Because of the particular architecture, Internet was conceived to be decentralized, borderless, and nearly unlimited in data capacity, all of which conflicted with the territorially-based sovereigns of power in the real world to regulate online activities. In their views, self-governance was not only the preferable and effective way in creating a cyberspace sought to be preserved, but also the legitimate process in cyberspace where traditional sovereign authority and external control were invalid.
Following the idea of Internet exceptionalism, the broad immunity or limited liability for Internet intermediaries were established, creating a legal environment where the norms and regulatory mechanism in the real world were effectively inapplicable. To govern the online relationship and approach intermediary liability for wrongdoing, a “Good Samaritan” defense was built into the CDA. Through this defense Internet intermediaries were welcomed for self-regulation but not duty-bound to stop illegal online content. The proposition of self-governance is also supported by arguing that civil liability is fault-based, therefore it was unfair to impose liability to intermediaries for illegal content when they have very limited awareness of the substance of users’ communications or transactions.
The three factors mentioned above supported the legal framework providing immunity or safe harbor to Internet intermediaries for the illegal content originated with third-parties. Additionally, these three factors are interconnected and cannot be analyzed separately. Free speech concern restrained government intervention, which promoted innovation and neutrality. Innovation boosted the prosperity of Internet which was helpful for free speech. The neutral position of Internet intermediaries stimulated end-user innovation and avoided collateral censorship.
Given the relative small and weak position of the Internet industry twenty years ago, it was reasonable to adopt these policies to promote its development. However, when the information environment has greatly changed and Internet intermediaries have risen to be the chokepoints of communication, it’s time to reexamine whether the rationales are still valid and the current legal framework is sufficient to realize those goals.
III. Are They Still Valid?
In Fair Housing Council of San Fernando Valley v Roommates.com, the majority of Ninth Circuit decided Roommates.com “materially contributed” to the online contents by requiring users to answer questions, which made it claim no immunity from Section 230 of CDA. In the EU, the Court of Justice of European Union ruled in L’Oreal v eBay, that eBay had played an “active role” in producing online contents which failed it to be exempted from liability. The two cases revealed the complex role of Internet intermediaries play when they are producing and displaying the online content. In contrast to the fragile means of communication that they used to be, Internet intermediaries have grown to be the chokepoints of communication.
Actually, this is only one of the big changes to Internet intermediaries. As the Internet pervades and Internet intermediaries’ business models evolve, they gradually become centralized platforms. They not only passively mediate information between parties, but also actively intervene the information environment where people live with. Given these changes, it is important to ask the question: are the rationales supporting the current legal framework which provided broad Internet intermediaries immunity or “safe harbor” still valid?
A. Internet Intermediary Becomes the Potential Threat to Free Speech
Existing rationale supporting the Internet intermediary immunity or “safe harbor” concerned the collateral censorship problem. Unlike traditional intermediaries, the communication model of Internet intermediaries shifted from a centralized “one-to-many” structure to a decentralized “many-to-many” structure. It even seemed as if the communication was conducted without intermediaries. However, as the role of Internet intermediaries grew, the rationale gradually became invalid.
According to Own Fiss, “the purpose of the First Amendment was to broaden the scope of public discussion to make people understand different opinions, which empower them to pursue their goals freely.” Similar ideas were also clearly expressed in Jerome Barron’s classic article. He strongly stated that “what matters is providing citizens greater access to conflicting viewpoints, not because speakers with disruptive ideas have rights to be heard, but because we as a society have an interest in hearing them.” Barron’s concern focused on the power of private censorship, especially the traditional intermediaries who controlled the access to mass media, which permitted only millionaires to speak while repressing competing ideas from others. In contrast to their traditional counterparts, like newspapers and broadcasters, Internet intermediaries offered a ubiquitous and comprehensive information environment where “not only everyone is capable to be heard, but also everything worth saying shall be said.” It was for this reason that Internet intermediaries were offered the privileges.
With the rapid development of technology and business model, however, the modern Internet is no longer merely an intermediary for “many-to-many” communication, but a dominant platform controlling the bottleneck of mass communication. Internet service providers were found to discriminate data and favor certain content or applications either by giving them different levels of priority or charging them differently. Search engines were frequently criticized for skewing search results to tip their own services. Participative network platforms such as YouTube and Facebook were also accused of manipulating online contents to benefit advertisers and investors. All of these comments showed growing concern for the power of the Internet intermediaries effect on the information environment. As Berman & Weitzner stated, “decentralized open access and user control over content are two key features to best serve First Amendment values relating to the freedom of expression.” First Amendment values were impeded when the Internet intermediaries became more closed, which held the communication bottleneck, and users could not control the content in cyberspace. . Therefore, if the Internet intermediaries become similar to their traditional counterparts who are likely to determine what the audience will hear and in what manner they will be able to respond, why should we treat them differently and offer Internet intermediaries extra privileges to be immune from liability?
Besides the broad immunity or “safe harbor” offered to Internet intermediaries, the current legal framework further restrained government intervention, encouraging self-regulation by intermediaries themselves. For example, the EU forced governments to require Internet intermediaries “to monitor the information which they transmit or store, not a general obligation actively to seek facts or circumstances indicating illegal activity.” However, the lack of government intervention doesn’t necessarily mean that there won’t be censorship any more. On the contrary, surveillance on online data is a ubiquitous business model which has already been adopted by nearly every Internet intermediary. For example, by tracking individuals’ searching history, search engines are able to push targeted advertisements, which are their main revenue sources. As Balkin stated, Internet intermediaries, who once claimed to be the democratized digital infrastructure of speech, also become “the infrastructure of surveillance and speech regulation.”
Despite the best efforts of great scholars to expand the scope of the First Amendment, the Supreme Court squarely foreclosed the possibility. Under current law, it is the government’s’ attempt to restrict freedom of speech that would be constitutionally problematic. Private power over free speech is still out of the reach of the First Amendment which cannot be invoked to require governmental intervention. However, as the impacts of Internet intermediaries on the information environment grow, it is still an open question whether more regulation should be imposed on them, especially when they have already moved away from the initial expectations of cyber-libertarians.
B. Innovation is Impeded: Collusion Among Oligarch
In the comments on Julie Cohen’s book Configuring the Networked Self, Balkin stated that “gaps in legal and technological enforcement might benefit the powerful far more than the powerless.” This also applies to the legal framework offering immunity and “safe harbor” for Internet Intermediaries. Although § 230 of CDA and § 512 of DMCA and other statutes grant the same privileges for all kinds of Internet intermediaries, it is the big player who benefits most. Worse still, when they gain power they will collude with other oligarchs to restrain latecomers and to impede innovation.
One example is the conflict between competition law and the privileges offered to Internet intermediaries by CDA and DMCA. In Search King v. Google, Google was sued for lowering the ranking of plaintiff’s websites on Google’s search results. The court denied the plaintiff’s application as it considered that Google’s search result was an opinion and thus entitled full protection of the First Amendment. According to the rule, Google was recognized to legitimately exercise an editorial decision and had a constitutionally protected right to choose what information to display in search results and what to exclude. However, when considering issues related to intermediary liability, like in Google France, Google Inc. v Louis Vuitton Malletier and Obado v. Magedson,et al., both courts concluded that Google would be immunized from liability as it was sufficiently neutral “in the sense that its conduct is merely technical, automatic and passive.” Thus, paradoxically, the defense that immunized Google in a competition case would eventually make it liable in an infringement case, basically concerning the same behavior. If we treat the search results as opinions of Google, it is weird to say that they are merely generated by an algorithm which is not based on purposeful action of Google.
YouTube Content ID system is another example to show how these Internet oligarchs affect the market and innovation. Content ID system is a suite of content management tools to give rights holders control of their content. When a video is uploaded, the system will search for a match in the Content ID database. In this way, rights holders can identify user-uploaded videos comprised entirely or partially of their content. Once a match is found, rights holders can choose, in advance, how to deal with them: license the materials to make money or block the content from YouTube altogether. Content ID system was fiercely criticized for failing to protect fair use and automatically blocking contents without transparency and due process.
Besides the YouTube Content ID system, we are seeing more similar cases about the collusion between Internet intermediaries and content industry. For example, Google and book publishers signed agreement for sharing online revenues related to the display of scanned books. Additionally, the Memorandum of Understanding (MOU) between several ISPs and copyright holders stated clearly their cooperation to take measures against infringing users. By collusion with right holders, Internet intermediaries might not only generate large profits by revenue sharing, but also squeezed out latecomers who might be a potential threat to the incumbents. After growing and becoming oligarchs, the Internet intermediaries cynically started to advocate to strengthen infringement liability for that which they were used to be relieved. This was similar to the theory of Ha-Joon Chang on the economic history, who argued that the developed countries would always “kick away the ladder” after they climbed up the hill and became rich while the developing countries were still struggling for growth.
Once promoted as a stimulus for innovation, the immunity regime for Internet intermediaries might turn in the opposite direction. It is not saying that legal framework is disadvantageous to small players, but rather a laissez-faire regulation system will finally lead to oligopoly and impede innovation.
C. Internet Intermediary Is Not the Exception
As noted above, Internet exceptionalism was based on the belief that Internet architecture was naturally neutral and decentralized. However, as Lessig convincingly argued, architecture itself was remarkably fluid and thus configurable. As he clearly noted:
Cyberspace…has different architectures…An extraordinary amount of control can be built into the environment…What data can be collected, what anonymity is possible, what access is granted, what speech will be heard—all these are choices, not “facts”. All these are designed, not found.
Building on the observation, Lessig warned that the greatest threat to the exceptional characteristics of cyberspace came from the perfect control embedded in the market forces. The “invisible hand”, he argued, had no motivation to protect the fundamental values promoted by cyber-libertarians, for which political and collective action was needed to counteract the influences of concentrated private power.
Lessig’s concern is not a theoretical hypothesis, but rather what is actually, currently happening. After twenty years since the current legal framework was enacted, several important changes have happened to Internet intermediaries and there was an increasing trend of converging control over access, content and users.
First, technological advancement and the greatly reduced cost of computing and storing data made it possible for Internet intermediaries to monitor all the content uploaded by users. The expert reports on Scarlet v SABAM stated thirteen feasible filtering systems, seven of which were considered possible to be deployed to filter P2P transmissions. The YouTube Content ID system is another example. Once considered to be impractical, it is common to see the surveillance system embedded in all kinds of Internet intermediaries.
Second, Internet intermediaries’ business model heavily rely on exploiting data on their platforms to make profits. Different from the cheap conduits they were used to be, Internet intermediaries have complicated their neutral role by getting involved into the process of content production and distribution. This was especially explicit in the L’Oreal v eBay case when eBay profited from advertising the infringing trademarks.
Third, more communication controlling systems are installed by Internet intermediaries, voluntarily or pushed by governments. Websites blocking and generated response are becoming popular mechanisms to fight against online infringements, which proposed serious problems of fundamental rights emphasizing equal and free access to the Internet.
As Internet intermediaries are more proactive in accumulating and distributing online content, the utopian expectations of cyber-libertarian fail, as well as the rationale supporting Internet intermediary immunity or “safe harbor”.
During the past twenty years, the immunity and “safe harbor” protection provided for Internet intermediaries against liability for online content that originated from third-parties had greatly promoted free speech and innovation. §230 of CDA, §512 of DMCA as well as the articles in the E-commerce Directive were considered to be the fundamental legal frameworks to facilitate a prosperous and robust Internet. After twenty years’ development, Internet intermediaries are no longer the fragile means of communication, but rather a dominant platform holding the bottleneck. Therefore, whether the existing legal frameworks are still valid may be questionable.
Three factors that supported the rationales of providing immunity or “safe harbor” to Internet intermediaries are examined while three corresponding critics are argued. Internet intermediaries themselves might harm free speech when they purposefully modified the information environment for commercial reasons. Collusion among monopolistic Internet intermediaries and powerful content industry might impede innovation and squeeze out late-comers. The centralized and non-neutral position Internet intermediaries held might not meet the expectation of Internet exceptionalism.
Despite these transformations, I’m not arguing here for direct government intervention and imposing strict liability on Internet intermediaries. On the contrary, this would merely replace one “big brother” for another and in turn deteriorate the current information environment we are living with. What is really needed is a balance between strict liability and complete immunity. We cannot afford another policy choice going to the other extremity, especially when Internet has pervaded the whole society.Further detailed policy suggestions go beyond this article. However, before we adopt any reform, a clear understanding of the rationales of the current legal framework and the reasons why they need modification is very important, which is exactly what this paper wants to achieve.
 Fulbright Scholar, California International Law Center, University of California, Davis; Ph.D. candidate, the School of Public Policy and Management, Tsinghua University, Beijing, China. This article is based on a working paper on American Society of Comparative Law Younger Comparativists Committee Second Workshop on Comparative Business and Financial Law at the Law School of UC Davis. I am grateful for Professor Anupam Chander for the valuable discussion and direction. I’m indebted to Professor John Hunt for comprehensive modification and revision. I thank Professor Frank Gevurtz for the newest materials offered to me from the Comparative Law Congress. I also thank Professor Afra Afsharipour, Professor Virginia Harper Ho, Uyen P. Le and Cathy Hwang for the valuable comments and encouragements. All errors are solely mine.
 Michael Nunez, “Former Facebook Workers: We Routinely Suppressed Conservative News”, Gizmodo (May 09, 2016), available at http://gizmodo.com/former-facebook-workers-we-routinely-suppressed-conser-1775461006 (Gizmodo, the technology news site, cited two former Facebook news curators as saying that Facebook routinely suppressed conservative content on the social network’s influential “trending” news. This report later aroused heavy critics against Facebook about whether we should continue to trust the neutral position of Internet intermediary and what liability should be imposed on them).
 Doe v. Internet Brands, Inc., 767 F.3d 894 (9th Cir. 2014). The plaintiff Jane Doe sued Internet Brands, Inc., the owner of Model Mayhem which was a networking website allowing third-parties to provide job information for models, alleging that they should be liable for the unlawful acts of others as they failed to warn her the potential risk on the website. Jane was lured to a house and then assaulted by two unrelated individuals through the website. The district court dismissed the plaintiff’s claim on the grounds that Internet Brands was immune from prosecution under CDA, 47 U.S.C. Section 230(c) (2012). However, the Ninth Circuit overthrew the judge, ruling that a tort based on duty that would require such a warning fell outside of section 230(c) and thus the website should be liable. The Court concluded that, “the CDA does not declare ‘a general immunity from liability deriving from third party content’” and “Congress has not provided an all-purpose-get-out-of-jail-free card for Internet intermediaries.”
 Delfi AS v. Estonia, (Application no. 64569/09) (2013). The case concerns the liability of an online news portal for the offensive comments posted by its readers below one of its online news articles. In Jun. 2015, the Grand Chamber of the European Court of Human Rights supported the Estonian courts’ decision and ruled that Delfi, the largest internet news portals in Estonia, was liable for the insult comments against plaintiff posted by readers, although Delfi had expeditiously removed them as soon as it had been informed. According to the court, as a professionally managed news platform who had an economic interest advertising income by inviting readers to post comments, Delfi was considered to be a provider of content services, rather than of technical services. As a publisher, Delfi was expected to take special care in assessing the potential risk of a specific article and effectively prevented clearly unlawful comments from being published.
 Fair Hous. Council of San Fernando Valley v. Roommates.com, LLC, 521 F.3d 1157 (9th Cir. 2008). Roommates.com is a website that operates to match individuals renting rooms with those who need rooms. In this case, the Ninth Circuit held that Roommates.com could not claim immunity under CDA § 230 where as a condition of use, as it required users to choose among set answers to questions, such as sexual orientation, family status of a future roommate, which violated the anti-discrimination laws. The court reasoned, by requiring subscribers to provide the information as a condition of accessing its service, and by providing a limited set of pre-populated answers, Roommate becomes much more than a passive transmitter of information provided by others. Therefore, it at least in part developed the content which exempted itself from the protection of section 230.
 Zeran v. America Online, Inc., 129 F.3d 327 (4th Cir. 1997). In this case, the plaintiff, Ken Zeran, was defamed by an unknown AOL subscriber who made several posting on AOL advertising that Zeran had for sale certain tasteless t-shirts regarding the bombing of the Alfred P. Murrah Building in Oklahoma City. Zeran was then inundated with telephone complaints and death threats. Although AOL removed the postings and cancelled the account of the unknown poster, Zeran still sued AOL for defamation. Zeran claimed that § 230 immunity eliminates only publisher liability, leaving distributor liability intact. According to Zeran, interactive computer service providers like AOL were normally considered instead to be distributors, like traditional news vendors or book sellers. However, the Court refused his claim, stating that “distributor liability is merely a subset, or a species, of publisher liability, and is therefore also foreclosed by § 230”.
 Most commentators strongly criticized the Zeran court’s conclusion, stating that could not be what Congress intended. See, e.g., Emily K. Fritts, Note, Internet Libel and the Communications Decency Act: How the Courts Erroneously Interpreted Congressional Intent with Regard to Liability of Internet Service Providers, Ky. LJ, 2004, 93:765. (arguing that “the First Amendment has never garnered an absolute right for the people to say whatever they want wherever they want”, therefore, “Congress should step in with a clearer mandate for the courts to keep in line with the traditional common law of defamation”); Robert T. Langdon, Note, The Communications Decency Act § 230: Make Sense? Or Nonsense? -A Private Person’s Inability to Recover if Defamed in Cyberspace, John’s L. Rev. 73 (1999): 829. (explaining “the elements that a person must prove to recover in a defamation action” and arguing that “[T]he Communications Decency Act impedes a private person’s ability to recover if defamed on the Internet by foreclosing adequate legal remedies.”); Sewali K. Patel, Note, Immunizing Internet Service Providers from Third-Party Internet Defamation Claims: How Far Should Courts Go? Vand. L. Rev. 55 (2002): 647. (concerning “[the]CDA immunity become an absolute bar to a plaintiff’s recovery” and arguing that “distributor liability is consistent with § 230(c) and therefore should be imposed on ISP”.
 See Doug Lichtman and Eric Posner, Holding Internet Service Providers Accountable, Sup. Ct. Econ. Rev. 221-59 (2006) (challenging “the trend of in the courts and Congress away from liability and toward complete immunity for Internet service providers” and arguing “service providers should bear some responsibility not only for stopping malicious code, but also for helping to identify individuals who originate it.”). See also Ronald J. Mann and Seth R. Belzley, The Promise of Internet Intermediary Liability, Wm. & Mary L. Rev. 47, 239 (2005) (arguing that “[T]he Internet’s rise has brought about three changes that make intermediaries more likely to be least cost avoiders in the Internet context than they previously have been in offline contexts.”).
 See Neal Kumar Katyal, Criminal Law in Cyberspace, U. Pa. L. Rev. 149.4, 1003-114 (2001) (arguing that “Internet service providers will often be essential in preventing cybercrime.”). See also Joel R. Reidenberg, States and Internet Enforcement, U. Ottawa L. & Tech. J. 1-18 (2004) (commending that “intermediaries offer the most efficient and attractive means to reach rule violators,” thus “states must find ways to transpose the powers of enforcement to the internet.”). See also Uta Kohl, The Rise and Rise of Online Intermediaries in the Governance of the Internet and Beyond–Connectivity Intermediaries, Int’l Rev. of L. Computers & Tech. 26.2-3, 185-210 (2012) (arguing that “intermediaries acting as transactional and communicative chokepoints have become popular regulatory targets.”).
 In this article, we mainly focus on speech intermediaries instead of other mediators like banks in money markets and brokers in real estate markets. For a detailed definition, see Thomas F. Cotter, Some Observations on the Law and Economics of Intermediaries, Mich. St. L. Rev. 67 (2006).
 See Thomas F. Cotter, Memes and Copyright, Tul. L. Rev. 80, 331 (2005).
 A more comprehensive analysis related to Internet intermediaries refer to the report of the OECD’s Committee for Information, Computer and Communications Policy (ICCP). In the OECD meeting at Ministerial level in Seoul in 2008, [T]he Declaration on The Future of the Internet Economy invited the OECD to examine “the role of various actors, including intermediaries, in meeting policy goals for the Internet economy in areas such as combating threats to the security and stability of the internet, enabling cross-border exchange, and broadening access to information”. In response, ICCP conducted the report, The Role of Internet Intermediaries in Advancing Public Policy Objectives, in which the legal liabilities problems of Internet intermediaries were proposed, available at http://www.oecd.org/sti/ieconomy/theroleofinternetintermediariesinadvancingpublicpolicyobjectives.htm.
 When it comes to matters of public figures, the plaintiff must also show some form of fault on the part of the defendant. See New York Times Co. v. Sullivan, 376 U.S. 254 (1964).
 See Anderson v. New York Tel. Co., 35 N.Y.2d 746, 320 N.E.2d 647 (1974). (Judge Gabrielli, J. concurring that “if there was no publication by defendant, then there is no need even to consider the further privilege and constitutional points also relied upon by the dissent.”).
 Restatement (Second) of Torts § 581 (1977).
 Smith v. California, 361 U.S. 147 (1959). Smith, a bookstore proprietor, was convicted of violating a Los Angeles City ordinance for having an obscene book in his inventory. Judicial interpretations of the ordinance made simple possession of obscene books unlawful even if the person possessing them had no knowledge of their contents. However, the Supreme Court reversed the state court and concluded that Mr. Smith was protected under the Due Process clause of the Fourteenth Amendment.
 Id. at 153-54.
 Knowledge of falsity is a kind of actual malice, which is a standard notoriously difficult to meet.
 See Yonchai Benkler, The Wealth of Networks: How Social Production Transforms Markets and Freedom (2006).
 See Karine Perset, The Economic and Social Role of Internet Intermediaries (2010).
 Cubby Inc. v. CompuServe Inc., 776 F. Supp. 135 (S.D.N.Y. 1991). CompuServe was an Internet service provider hosting an online news forum where third-parties could upload contents for subscribers. Cubby Inc. sued CompuServe for defamatory contents in its forum. The court rejected the plaintiff’s claim and ruled that CompuServe was merely a distributor, rather than a publisher. As a distributor, CompuServe could only be held liable for defamation if it knew, or had reason to know, of the defamatory nature of the content. The case established a precedent for Internet service provider liability by applying defamation law, which was originally intended for hard copies of written works, to Internet intermediaries.
 Stratton Oakmont Inc. v. Prodigy Services Co., 1995 WL 323710 (N.Y. Sup. Ct. 1995). Prodigy was an online service provider running a bulletin board on which users could upload contents. Stratton Oakmont, Inc. sued Prodigy for the defamatory contents in the bulletin board from an unidentified user. The court held that Prodigy was liable as the publisher of the content created by its users because it exercised editorial control over the messages on their bulletin boards. The court found that Prodigy’s decision to implement a screening program and to allow its Board Leaders to remove content evidenced sufficient editorial control to deem Prodigy a publisher.
 Although the Stratton Oakmont court distinguished itself from Cubby by two points, it may not be sufficient to support an exactly opposite judgment. The Stratton stated that, “First, Prodigy held itself out to the public and its members as controlling the content of its computer bulletin boards. Second, Prodigy implemented the control through its automatic software screening program.” In this way Prodigy was considered to be publisher rather than distributor. Despite differences in business strategy, Prodigy and Compuserve’s roles as Internet intermediaries were the same. The Cubby and Stratton decisions meant that an ISP that attempted to monitor content on its network would most likely be subject to publisher liability, while an ISP that followed a completely “hands-off” approach would only be subject to distributor liability. See also Freiwald, Susan. “Comparative Institutional Analysis in Cyberspace: The Case of Intermediary Liability for Defamation.” Harv. JL & Tech. 14 (2000): 569.
 Religious Technology Center v. Netcom Online Communication Services, Inc., 907 F.Supp. 1361 (N.D. Cal. 1995). In this case the plaintiff, Religious Technology Center (“RTC”), argued that defendant Netcom, the operator of a computer bulletin board service (“BBS”), was directly, contributorily, and vicariously liable for copyright infringement. The court ruled that Netcom could not be held directly liable for any infringing material posted by its clients since Netcom itself did not upload the material. As RTC could prove that Netcom had knowledge of the infringing activities, the contributory infringement claim of the plaintiff was also refused. The court finally concluded that RTC’s claims of direct and vicarious infringement failed. This case set an important precedent stating that any interceding service that did not upload infringing content directly should not be held directly responsible for the illegal and/or infringing actions of its customers.
 Playboy Enterprises Inc. v Frena, 839 F. Supp. 1552 (M.D. Fla. 1993). In this case, the plaintiff, Playboy Entertainment, charged the defendant, George Frena, who operated a subscription computer bulletin board service (BBS) for copyright infringement. The defendant claimed to have never uploaded Playboy’s copyrighted photographs which were posted by subscribers and as soon as he was acknowledged about the infringements, he removed the photographs. However, the court held that it did not matter that Frena was not the originator of the authorized copies, because as long as he supplied a product containing unauthorized copies, he violated the plaintiff’s exclusive rights. Additionally, it didn’t matter that the defendant might have been unaware of the copyright infringement; intent or knowledge are not an elements of infringement, and thus even an innocent infringer can be liable for infringement. See Sega Enterprises Ltd. v. MAPHIA, 948 F. Supp. 923 (N.D. Cal 1996).
 Id. at 1562.
 Common Decency Act, 47 U.S.C. § 230(e)(1) (1996).
__ 47 U.S.C. § 230(c)(1)
 47 U.S.C. § 230(c)(1)
 47 U.S.C. § 230(e) (2).
 Digital Millennium Copyright Act, 17 U.S.C. § 512 (1998).
 The reason why copyright was given different treatment from other infringements lies in the systematic bias within the legislative process. Copyright scholarship argued that the well-organized group of copyright owners had used their political power to affect the legislation while the public was unable to effectively advocate for themselves. Therefore, the protection of copyright infringements is better than protection of defamation. See also Lev-Aretz, Yafit. “Copyright Lawmaking and Public Choice: From Legislative Battles to Private Ordering.” Harv. 27. JL & Tech. 203 (2013).
 17 U.S.C. § 512.
Council Directive 2000/31/2000 O.J. (L 178) 1 (EC).
 See Kohl U. The rise and fall of online intermediaries in the governance of the Internet and beyond–connectivity intermediaries. International Review of Law, Computers & Technology, 2012, 26(2-3): 185-210.
 See Miguel Peguera, The DMCA Safe Harbors and Their European Counterparts: A Comparative Analysis of Some Common Problems, 32 Colum. J..L. & Arts 481 (2009).
 See James Grimmelmann,. “Internet Law: Cases and Problems 4.0.” 177-189 (2014)..
 Berin Szoka, “Section 230: The Cornerstone of Internet Freedom, Technology Liberation Front, (Aug. 18, 2009), https://techliberation.com/2009/08/18/section-230-the-cornerstone-of-internet-freedom
 Abby Ohlhelser, Revenge porn purveyor Hunter Moore is sentenced to prison,” The Washinton Post: The Intersect, (Dec. 3, 2015), https://www.washingtonpost.com/news/the-intersect/wp/2015/12/03/revenge-porn-purveyor-hunter-moore-is-sentenced-to-prison
 Suzanne Choney, “Classified ad site Backpage in crosshairs over child sex ads,” NBC News: Tech News, (July 29, 2013), http://www.nbcnews.com/tech/tech-news/classified-ad-site-backpage-crosshairs-over-child-sex-ads-f6C10789250.
 Joel R. Reidenberg, et al., Legal Research Paper, Section 230 of the Communications Decency Act: A Survey of the Legal Literature and Reform Proposals, Fordham L. Sch.—Ctr. on L. & Info. Pol’y 23-24, 46 (2012) (also noting that “the majority of scholarly literature identified is critical of section 230”).
 Marcelo Thompson, Beyond Gatekeeping: The Normative Responsibility of Internet Intermediaries, Vand. J. Ent. & Tech. L. 4 (2015).
 Michael L. Rustad & Thomas H. Koenig, Rebooting Cybertort Law, 80 Wash. L. Rev. 335, 376 (2005).
 The relationship between Internet and free speech is bidirectional. On one hand, Internet promotes free speech. On the other hand, the notion that free speech asis a fundamental right also supports the development of Internet. Anupam Chander &Uyen P. Le, Free Speech, 100 Iowa L. Rev. (2014) 501, 504. In this section, I mainly focus on the positive impacts of Internet on free speech and its prerequisite to be realized.
 Jack M. Balkin, Free Speech and Hostile Environments, Colum. L. Rev..,. 2295-2320 (1999).
 Id. at 2298.
 Frederick Schauer, Fear, Risk and the First Amendment: Unraveling the Chilling Effect, 58 B.U. L. Rev. 685, (1978).
 Seth F. Kreimer, Censorship by Proxy: The First Amendment, Internet Intermediaries, and the Problem of the Weakest Link, 155 Penn St. L. Rev. 11, (2006).
 Eugene Volokh, Cheap Speech and What It Will Do,speech and what it will do. 104.7 Yale L.J. 1805, (1995).
 Reno v. ACLU, 521 U.S. 844 (1997). (In this case, Supreme Court struck down the anti-indecency provisions of the Communications Decency Act (CDA), stating that they violated the First Amendment’s guarantee of freedom of speech. The Supreme Court explained that the Government may not “reduc[e] the adult population…to…only what is fit for children”. This was the first major Supreme Court ruling on the regulation of materials distributed via the Internet. Although the anti-indecency portions of the CDA were ruled unconstitutional, section 230 survived and became “the most important piece of law” in cyberspace); See James Grimmelmann, Internet Law: Cases and Problems 4.0, 200 (2014), http://internetcasebook.com/
 See Seth F. Kreimer, Censorship by proxy: the First Amendment, Internet Intermediaries, and the Problem of the Weakest Link. U. Pa. L. Rev. 11-101. (2006) (explaining more comprehensively about the collateral damage of proxy censorship).
 Zeran v. America Online, Inc., 129 F.3d 327 (4th Cir. 1997). See Flex T. Wu, Collateral Censorship and the Limits of Intermediary Immunity, 87.1 Notre Dame L. Rev. 293, (2011).
 Julie E. Cohen, Configuring the Networked Self: Law, Code, and the Play of Everyday Practice, 234 (2012).
 Id. at 246.
 Jack M. Balkin, Room for Maneuver: Julie Cohen’s Theory of Freedom in the Information State, 6 Jerusalem Review of Legal Studies 92, 79-95 (2012).
 Reinier Kraakman laid the legal foundation of gatekeeping theory, describing that how regulators can make use of gatekeepers’ privileged positions for law enforcements. See also Reinier H. Kraakman, Gatekeepers: the Anatomy of a Third-party Enforcement Strategy, 2 J.L. Econ. & Org. 53, 104 (1986).
 See Jonathan L. Zittrain, The Generative Internet, 119 Harv. L. Rev. 1974, 2040 (2006). See also Jonathan L. Zittrain, The Future of the Internet and How to Stop It (2008).
 See Jonathan L. Zittrain, The Future of the Internet and How to Stop It 261 (2008).
 See Anupam Chander, How Law Made Silicon Valley, 63 Emory L.J. 645, 639-694 (2013).
 Internet exceptionalism was an influential theory in the early days of Internet, stating thatthe online world was naturally independent of sovereignty power in the real world. It refused regulation extended from physical world, arguing that cyberspace would develop its own effective legal institutions. Leading scholars proposing Internet exceptionalism including David R. Johnson, David Post and John Perry Barlow. See Johnson, David R., and David Post. Law and borders: The rise of law in cyberspace. Stanford Law Review (1996): 1367-1402. See also Barlow, John Perry. A Declaration of the Independence of Cyberspace. (1996), available at http://wac.colostate.edu/rhetnet/barlow/barlow_declaration.html
 See Radin, Margaret Jane, and R. Polk Wagner. The myth of private ordering: rediscovering legal realism in cyberspace. available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=162488, (proposing the word “cyber-liberatarians”).
 The Internet architecture was called “end-to-end”, which contemplates networks designed so that intelligence rests in the ends, and the network itself remains simple. See Lawrence Lessig, The Architecture of Innovation, 51 DUKE L.J. 1783, 1789 (2002).
 See Isenberg, David. “Rise of the stupid network.” Computer Telephony 5.8 (1997): 16-26. (arguing that the value of network is based on intelligent end users which must be supported by a stupid network).
 521 F.3d 1157 (9th Cir. 2008).
 L’Oreal SA & Ors v EBay International AG & Ors  EWHC 1094. The case concerned eBay’s liability for the sale of counterfeit L’Oréal products on its UK Web site. One of Oreal’s important claims was eBay should be primarily liable for the use of keywords in the Link Mark and sponsored links on third party search engines, both of which attracted customers and directed them to those infringing goods. The court ruled that, the use of keywords by eBay was not only to promote its own services as an online marketplace, but also to promote its users’ postings which helped attract customers. When eBay provided assistance like optimizing the presentation of online offers for sale or promoting those offers, it played an “active role” which gave it knowledge of or control over the data relating to the offers for sale. Although E-Commerce Directive restricted the intermediary liability, it only applied where that operator had not played an active role allowing it to have knowledge or control of the data stored. Therefore, eBay was held liable by the court.
 See Sullivan, Kathleen M. First Amendment Intermediaries in the Age of Cyberspace. 45 UCLA L. Rev. 1653, (1997) (arguing that “the decentralization of speaking and listening over the Internet eliminates a host of familiar middleman”).
 See Fiss, Owen. The Irony of Free Speech. Harvard University Press (2009).
 See Barron, Jerome A. Access to The Press. A New First Amendment Right. Harv. L. Rev. 1641-1678, (1967).
 See Meiklejohn, Alexander. Political Freedom (1965).
 See Formal Complaint of Free Press and Public Knowledge against Comcast Corporation for Secretly Degrading Peer-to-Peer Applications, Memorandum Opinion and Order, WC Docket No. 07-52 (2010).
 See Oren Bracha & Frank Pasquale, Federal Search Commission? Access, Fairness, and Accountability in the Law of Search, 93 CORNELL L. REV. 1149, 1161–79 (2008). (arguing that “search engines are capable to manipulate and structure the search results…and neither market discipline nor technological advance is likely to stop it”). See also Jennifer A. Chandler, A Right to Reach an Audience: An Approach to Intermediary Bias on the Internet, 35 HOFSTRA L. REV. 1095 (2007). (arguing that “despite the positive stories, there are also numerous stories of how intermediaries like search engines undermine the flow of information from speaker to listener.”)
 See Rebecca Tushnet, Power Without Responsibility: Intermediaries and the First Amendment, 76 GEO. WASH. L. REV. 986, 996–1002 (2008). (arguing that “Facebook and Youtube are likely to sacrifice individual users for a better image for advertisers and investors”).
 See Jerry Berman & Daniel J. Weitzner, Abundance and User Control: Renewing the Democratic Heart of the First Amendment in the Age of Interactive Media, 104 YALE L.J.1619, 1628–29, 1636–37 (1995)
 Guadamuz, Andrés. “Developments in Intermediary Liability.” Research Handbook On EU Internet Law, Andrej Savin and Jan Trzaskowski (eds), Edward Elgar (2014) (2013).
 See Balkin, Jack M. “Old school/new school speech regulation.” Harvard Law Review, Forthcoming (2014).
 There are still some objections arguing that the First Amendment does not prohibit government regulation on Internet intermediaries. See, e.g., Oren Bracha & Frank Pasquale, Federal Search Commission? Access, Fairness, and Accountability in the Law of Search, 93 CORNELL L. REV. 1149, 1161–79 (2008).
 Balkin, Jack M. “Room for maneuver: Julie Cohen’s theory of freedom in the information state.” Jerusalem Review of Legal Studies 6.1 (2012): 79-95.
 Search King v. Google, No Civ-02-1457-M, 11-12 (WD Okla 13 January 2003). The plaintiff sued Google for intentionally lowered the ranking of the plaintiff’s websites on Google’s search engine. The court rejected the plaintiff’s claim and upheld Google’s argument, stating that “[a] page rank is an opinion protected by the First Amendment, and any act aimed at knowingly and intentionally modifying the ranking of websites is legitimate expression of the freedom of speech.” Similar opinion can refer to Langdon v Google, Yahoo! and Microsoft Corp., 474 F Supp 2d 622 (D Del 2007).
 In Europe, Google was involved into a series of anti-trust litigations for lowering search results of its competitors, which was claimed to impede competition. See e.g., Analyzing Google’s Public Response to the EC’s Statement of Objections, available at http://www.foundem.co.uk/fmedia/Foundem_Jun_2015_Analysis/
 Google France, Google Inc. v Louis Vuitton Malletier, C-236/08 CJEU,Grand Chamber (23March 2010). The plaintiff sued Google for trade mark violations. According to the plaintiff, the search results of Google, under the heading “sponsored links”, linked to imitation versions of plaintiff’s products. The court rejected the plaintiff’s claim, stating that the search results were “natural results” and Google’s role is “neutral” which arguably put them beyond the liability.
 Obado v. Magedson, et al, 2014 WL 3778261 (D. N.J. July 31, 2014). The plaintiff sued several Internet intermediaries including Google for displaying allegedly defamatory information. The court rejected the plaintiff’s claim, stating that search results were determined by algorithm based on the content produced by third-part sites, and not by some purposeful act of search engines to create the content. Therefore, the defendants including Google were immune from liability.
 Google France (CJEU), ibid, para 114.
 See Gotham, Elizabeth, Lessons from Content ID: Searching for a Balance between Editorial Discretion and Free Expression on Application Platforms. Available at SSRN 2258861 (2012).
 See Publishers and Google Reach Settlement, available at http://publishers.org/news/publishers-and-google-reach-settlement (Last accessed June 26, 2016).
 See Memorandum of Understanding, Center for Copyright Understanding (July 6, 2011), available at www.copyrightinformation.org/wp-content/uploads/2013/02/Memorandum-of-Understanding.pdf (Last accessed June 26, 2016).
 See Op-Ed: When your YouTube video becomes a corporate profit center, available at http://www.latimes.com/opinion/op-ed/la-oe-0628-witt-youtube-copyright-20150628-story.html (Last accessed June 26, 2016).
 Ha-Joon Chang observed that the developing countries were forced by the developed countries to adopt a set of “good policies”, such as liberalization of trade and strong patent law, to foster their economic development. However, before the developed countries became rich, they usually preferred trade protectionism and poor patent protection. Ha-Joon Chang thus argued, the developed countries “kicked away the ladder” after they succeed and taught the developing country a contrary way. See Chang, Ha-Joon. “Kicking away the ladder.” (2002).
 See Lessig, Lawrence. Code and other laws of cyberspace. Vol. 3. New York: Basic books, 1999.
 Id. at 217.
 Id. at 6, (arguing that “the invisible hand, through commerce, is constructing an architecture that perfects control—an architecture that makes possible highly efficient regulation).
 Id. at 225-30 (discussing the need for greater and better democracy as a response to a changing cyberspace).
 Case C-70/10, Scarlet Extended SA v. Société belge des auteurs, compositeurs et éditeurs SCRL (SABAM), 2011 E.C.R. I-11959.
 Christina Angelopoulos, Filtering the Internet for Copyrighted Content in Europe, Amsterdam Law School Research Paper No. 2012-04, 9 (2012).
 See Twentieth Century Fox Film Corp & Ors v British Telecommunications Plc., 2011 EWHC 1981 (the plaintiff, several Hollywood film studios, sought to order British Telecommunications (BT), an Internet service provider, to filter content from Newzbin, a popular file-sharing site which was found copyright infringement in a former case. The High Court agreed with the plaintiff and issued an order to force BT technically block access from its subscribers to the Newzbin website.); Corynne Mcsherry, Graduated Response Program: Let’s Press the Reset Button (2012), https://www.eff.org/zh-hans/deeplinks/2012/04/graduate-response-program-lets-press-reset-button-backroom-deal.
 See Scarlet v. SABAM. (a conflict between the E-Commerce Directive (ECD) and the European Convention on Human Rights (ECHR) was explored by the European Court of Justice (ECJ). The plaintiff, the Belgian Society of Authors, Composers and Publishers (SABAM) sued against Scarlet, an Internet service provider for its users’ illegal downloading behavior. SABAM wanted Scarlet to install filtering software and curb further infringement. The court considered that there were fundamental rights concerning Internet access of users that would be affected by the filtering system proposed by SABAM. Therefore, the court rejected the plaintiff’s proposition. However, the court further stated that specific injunctions are still allowed).