Search Engine Law and Freedom of Expression. A European Perspective

Introduction¹ 

A typical exposition of the challenges to freedom of expression in the context of search engines starts with a discussion of Internet censorship in China and other countries that do not have a tradition of freedom of expression and a free press. As a result, the censorship of the Web and the suppression of search results by global search providers in China is well documented² and widely discussed in the media.³ The debate about the proper implications of freedom of expression for search engine law and policy in the European context has been much less prominent. This is regrettable because a variety of issues warrant such a discussion. Before addressing these issues in more depth, it is useful to start with a few examples.

Over the last two years, Martin Leguizamon, an Argentinian attorney, secured court orders for over a hundred people to get search results removed from Google and Yahoo.⁴ Some of the court orders related to searches about public officials. If anything, democratic governance and the freedom of expression imply the freedom to discuss and inform oneself about the functioning of public officials. This has led Google to publicly state its discomfort with Argentinian laws and appeal the respective court orders.⁵ Remarkably, in its plea for search engine freedom, Google refers to the protection it receives in Europe. Unfortunately, this reference is mistaken. In most European countries, search engine providers receive no clear legal protection against requests for removal of search results and struggle with similar issues as in Argentina.

In 2008, an artist and a photographer successfully sued Google in Germany for showing thumbnails of their works in its image search service.⁶ The material was posted online with the authors’ consent, but the court in Hamburg ruled that this does not imply that Google can make it searchable through an image search engine. The possibility for webmasters to exclude material from search engines by using the robots.txt instruction is also irrelevant, according to the court in Hamburg. The court concluded that search engines need express permission of copyrights holders, in the absence of which the use of thumbnails amounts to copyright infringement. Google has appealed the German ruling for obvious reasons. If the ruling stands, Google might have to shut down or restructure its image search service in Germany. Needless to add, it would also obstruct others than Google from providing image search services in Germany in the way that seems to have arisen as a standard on the Web.

In September 2007, a French court in Paris ordered Google to hand over the IP addresses of users of its search service to an AdWords customer, the French company Attractive Ltd.⁷ This developer of bars and restaurants in Paris contested the monthly Google bills for its advertisements on the Google page. An IP address with a time stamp is the key to the identification of Internet users. Although European regulators, amongst others, have started to scrutinize search engine providers for the systematic surveillance of users, it is likely they will continue to log a detailed set of non-anonymous data for lengthy periods of time. This information is available not only to the providers themselves, but, subject to conditions in national laws, also to third parties, including law enforcement and national security agencies. The effects of user surveillance on access to information and user behavior are not yet clear, but it is obvious that some users could start to self-censor themselves when considering using search engines for more controversial or sensitive subjects.

These three anecdotes clarify some of the typical challenges to freedom of expression and information in the context of Web search engines. In the following, I will give a systematic overview of these challenges and show that from the perspective of freedom of expression, which includes not only the freedom to express but also the freedom to seek and receive information and ideas, European search engine law and policy is far from ideal.

The importance of search engines

Search engines are amongst the most dominant media on the Web. Since the Web does not provide a search functionality itself, except for the limited navigational value of domain names, this gap has been filled by a variety of providers, tools and practices.⁸ Parallel to the growth of the Internet and the Web, there is an ever growing need for effective mediation between information sources and audiences. The informational abundance has made effective search tools extremely important and powerful. As first points of access to the Web, general purpose search engines have become the dominant navigational tool for Internet end-users. They have also become some of the most important platforms for information, knowledge and advertisements to find their way to an audience. Thus, the Internet and effective search tools in particular provide end-users with unprecedented levels of access to information. However, new modes of suppression of the relative accessibility of information have started to interfere with such access.

Pressure on entities controlling “accessibility”

There is a general recognition that traditional intervention or strategies focusing on the governance of the availability of content have become less effective. As a result some of this attention has shifted towards intermediaries that control the accessibility to content such as search engines and internet access providers. Here, availability means that the material exists somewhere on the Internet and accessibility stands for the relative “reachability” of available material.⁹ Among the reasons for this shift in focus are the abundance of information arising from the ease and variety of techniques for online publication, the global nature of the Internet with related jurisdictional problems, and the immaturity of enforcement mechanisms to tackle illegal material at its source.¹⁰

Internet service providers (ISPs) are amongst the primary targets to regulate the accessibility of content. Ofcom recently concluded that self-regulation by ISPs, which have strong ties with the jurisdictions of end-users, is a crucial element in a policy that aims to manage content risks for children.11 The discussion about the censorship of a Wikipedia entry in the United Kingdom is an illuminating example as well. The Internet Watch Foundation (IWF) placed the Wikipedia entry about a Scorpions cover on its list of child pornographic material that is used as a black list by major ISPs.¹² The discussion in the media that followed and the widespread availability and accessibility of the targeted image caused the IWF to review its decision and remove the Wikipedia entry from its list.¹³ There are various “voluntary” Internet filtering schemes directed at child pornography in Europe and the United States: quote-unquote voluntary because adopted under government pressure.¹⁴ More recently, rights holders have started to lobby for filtering and disconnection of users by ISPs in the context of illegal file sharing. It is important to note here that the possibility of circumventing blocking and filtering by Internet access providers is well-documented, so these measures can hardly be seen as preventing access to the blocked or filtered material entirely.

Like ISPs, search engines are being pushed or asked to help to regulate the accessibility of online information. In the United States, the Digital Millennium Copyright Act incentivizes search engines to remove results when receiving a notice that they link to copyright infringing material.¹⁵ All major U.S. based search engines react to such notices by removing the results globally. In Europe, a variety of laws put search engines under legal pressure to remove results from their index. In the Netherlands, Zoekmp3, a search engine for mp3-files, was found to be illegal altogether because of its exploitation of infringing material on the Web.¹⁶ Instead of using Zoekmp3 to target the source of the infringements, rights holders sued the service that made the material more readily accessible. Since 2004, major search engine providers in collaboration with government agencies in France and Germany remove some search results to illegal and harmful information from their country-specific services.¹⁷ Others have proposed to target the interface between users and search engines. For instance, European Commissioner Frattini called on search engines to make certain “dangerous” searches impossible.¹⁸ In all these cases the material remains available online and accessible through other means, for instance through the non-country-specific versions of the same search providers.

Freedom of expression

Without passing a definitive judgment on whether these developments are appropriate responses to what can in general be considered to be valid concerns, it is clear that freedom of expression theory should inform us about the legitimacy of this move to the suppression of accessibility of information. Whereas censorship and suppression of findability might be the logical outcome of an information environment that cannot be kept “clean” entirely, the same findability is essential for the freedom of expression in an environment that is characterized by information overload. It is not the ability to speak that is at risk as much as the ability to be found, and it is not the ability to access information that is at stake as much as the ability to locate that information. This brings us to the important question of what the right to freedom of expression actually protects.

Freedom of expression is protected as a fundamental right on the international level in United Nations Treaties,19 at the regional level in Article 10 of the European Convention on Human Rights (ECHR)²⁰ and at the national level in constitutions. Broadly defined, as in Article 10 ECHR, freedom of expression stands for the freedom to hold opinions and to receive and impart information and ideas without interference by public authorities and regardless of frontiers. It is informed by democratic ideals, the search for truth and individual self-fulfillment. ²¹ Notably, it does not only protect the freedom to speak, but also other communicative freedoms such as the freedom to gather, look for, access, and transmit information and ideas. The freedom to seek is explicitly mentioned in the UN treaties. Freedom of expression as protected in Article 10 ECHR is not absolute, but allows for certain narrowly defined interferences that are necessary in a democratic society.²² It is primarily a negative right, protecting private parties against state interferences such as censorship. However, freedom of expression can also place positive obligations on the state to guarantee the respect for the principle of freedom of expression in society and a minimum ability to effectively exercise one’s right to freedom of expression. Under Article 10 of the European Convention, the state is for instance ultimately responsible for a pluralist information environment.²³

Two types of legal pressure on search engines If one analyzes the pressure that is placed on search engines to suppress information in Europe, one can make a distinction between two types. First, there is pressure that is specifically directed at search engines and meant to suppress access to certain information. Second, there is pressure, more structural in nature, caused by the lack of acknowledgment of search engines’ importance for the online information environment. Both types of pressure are problematic from the perspective of freedom of expression. The first type can incentivize search engines to censor or obscure the presence of certain information on the Web, thereby obstructing legitimate information flows. The second type increases the difficulty of operating a search engines in Europe and prevents the development of a wide variety of effective services and tools that help to mediate effectively between information sources online and Internet users.

Liability of search engines for search engine results

When we look at the issue of censorship by search engines in Europe, a continuing problem is the absence of clear rules about the obligations of search engines when they are confronted with requests to remove certain results from their index.²⁴ A variety of national laws can make search results legally problematic, such as laws relating to copyright and trademark infringement, privacy, defamation, and certain forms of illegal speech such as hate speech or child pornography. Because of its massive automatic crawling of online information, a crawlerbased search engine is bound to end up indexing illegal information published by third parties. The question is what a search engine should do, when a government agency requests the removal of a search result,²⁵ when a TV-hostess asks for removal of results that connect her to adult websites,²⁶ or when the Church of Scientology asks it to remove results that allegedly infringe its copyright.²⁷ Should it remove or obscure these or some of these results or should it refuse to remove material without a valid legal order? Should it inform the particular website that it has been removed from the index and provide an opportunity to the website to contest the alleged illegality of their publications and to be restored in the index? Should it make it clear to users that results are being removed or obscured and for which reasons?

The legal regime for search engines in Europe does not at all guarantee that search engines will get the answers to these questions right. The current regime and the differences in legal rules between different jurisdictions have caused search engines to navigate between three options, when asked to remove material. They can adopt the lowest common denominator and remove information globally when there is a legal request for information, they can remove results from country-specific services, or they can refuse to comply with the request without a valid court order. The first option is the simplest, but results in the suppression of legitimate speech because of overbroad requests and cultural differences regarding tolerance for the publicity of information. The second option is in some ways more nuanced, but it does not prevent access to the information for the Internet savvy and results in the fragmentation of the Internet through intermediaries.²⁸ As the legal obligations are unclear, the third option is currently the best from the perspective of freedom of expression. But costly litigation, legal penalties, unfavorable legal precedent and regulatory pressure can make this option unattractive.

To summarize, European laws with regard to the liability of search engine providers for unlawful references do not protect users against facing undue suppression of search results. In addition, there is a significant lack of transparency about these practices. Since one can hardly expect major search engines to thoroughly investigate every complaint they receive, the existing uncertainty is bound to lead to the removal of legitimate websites, thereby diminishing their accessibility for end-users.²⁹ It is fair to add that third parties that feel that certain search results harm them in their legally protected interests find themselves struggling to understand the circumstances under which they can legitimately claim to get certain information removed from a search engine. Instead of addressing the issue, the European Commission and most European legislatures have postponed dealing with the liability of search engines for nearly a decade.³⁰

Legality of basic search engine operations

The basic operations of search engines are still the subject of litigation as well.³¹ Crawling, indexation, archiving, the showing of parts of the index to users, and linking continue to pose difficult legal questions. The reason is that a significant part of the Web consists of copyright protected works and the basic operations of search engines involve actions that may be restricted through copyright law.

Recently, the European Commission issued a Green Paper that touched upon the legality of basic search engine operations from the perspective of European copyright law.³² This Green Paper addresses the emergence of search engines as new successful commercial players in the online information environment and asks whether copyright law still guarantees the fair balance between rights holders and these new entities. The Green Paper seems to imply that search engines are not in full compliance with copyright law and are making their profits at the expense of rights holders, thereby increasing the pressure already placed on search engines through litigation. In a footnote, the Commission sceptically refers to one of the legal arguments search engine providers have used to claim that their basic operations are legal under copyright law. Search engines have claimed that since webmasters can opt out of indexation by a search engine by using a robot exclusion instruction such as robots.txt, webmasters implicitly give permission for the indexation of the material they publish online if they don’t. ³³

The Green Paper is an excellent illustration of the ambivalent attitude of European legislatures. On the one hand, they embrace the Web and the Internet as essential for the Information Society. On the other hand, they seem unwilling to embrace the new players that have naturally emerged to make the environment work. One cause for this ambivalence may be the increasingly hostile relationship between traditional publishers and search engines in Europe. Online publishers are still struggling to establish sustainable operations on the Web. They have to make material freely accessible to reach an audience through search engines, but see a large portion of advertising revenues go to others.34 Another explanation of this relatively unsympathetic attitude towards search engines might be found in the striking dominance of U.S. based companies in the online environment, and Google in particular.

Dominance of Google in Europe

Google’s market share in Europe, ranked by searches, is around 80 percent and continues to grow.³⁵ This dominance makes the way Google navigates the European legal waters particularly interesting. A website removed from Google practically ceases to exist and other search engines will be tempted to copy the practices of the leading search engine. More importantly, it raises the question of whether European law and policy is doing enough or could do more to facilitate the development of effective search tools and services and promote competition between search engines.

The government-funded Quaero project launched by France and Germany in 2006 has been portrayed as the European answer to Google.³⁶ The project soon split up into the Quaero and Theseus projects in France and Germany respectively. Both Quaero and Theseus intend to promote European search engine research and development, but the projects have not yet resulted in consumer search services with a notable user base. In particular, neither Quaero nor Theseus aims to safeguard freedom of expression and information in the search engine environment.

The point here is that instead of funding an industry that is developing incredibly fast without public spending, European legislatures could try to facilitate the search engine market by adopting laws and policies that are more favorable for new entrants. The legal uncertainty with regard to the liability and the basic operations of search engines constitutes an additional barrier to entry into the European search engine market. As this uncertainty results in significant legal overhead, a large established provider like Google is in the best position to deal with legal pressure, be it in the form of complaints, litigation, or from regulators. The more restrictive obligations are being placed on search engine providers, to negotiate with governments, to settle with publishers and to deal with complaints from third parties, the more difficult it gets for smaller providers to provide a similar service. In fact, it may be a remarkable achievement that Google, with the help of a small army of lawyers, has been able to keep its search services running in Europe.

Clearly, there is room and need for more than one general search engine, so European search engine law and policy should look beyond the dominant position of Google. The dominance of Google in combination with its central position in the ongoing legal debates about the proper responsibility of search engines and the legality of their basic operations could lead to a reinforcement of its dominance, especially if legislatures and regulators would impose additional legal obligations on search engines providers in Europe that are more difficult to fulfill for small providers.

Free access to information

Finally, if we address the freedom to search for users, it is important to reflect on search engines’ collection of unprecedented amounts of user data, and the subsequent use of these data by search engines and third parties. Systematic user surveillance has been widely discussed since the AOL data release.³⁷ Privacy advocacy groups, international data privacy bodies and in particular the EU’s Article 29 Working Party have reacted with recommendations that aim to minimize the processing of user data by search engines.³⁸

Central to the debate has been the retention term for search engine logs, which include the IP addresses of users, unique cookie data, search queries, timestamps, and user clicks. It is important to note that even if the retention terms for search engine logs were limited to the recommended 3-6 months, users of search engines would still be placed under systematic surveillance when using the key tools to navigate the Web. Since searching the Web is for many a daily activity, the search logs of a typical Internet user over 3-6 months provides an incredibly detailed picture of that user’s interests, activities, and intentions. If it is unfeasible to require search engine providers to minimize user surveillance further, the law should guarantee the confidentiality of search activity through other means.

User privacy in the context of search engines is primarily a concern related to free access to information for users. To guarantee the confidentiality of private communications and to protect the freedom of communication, communications law has traditionally placed special restrictions on the access to private communications by communications providers and third parties. The rules relating to access to the postal mail and wiretapping can serve as examples. Libraries also tend to highly value the privacy of their patrons to guarantee free access to information and ideas. In reality search activity is partly displacing the browsing through libraries and their catalogues, and it may be of an equally private nature as communications over the mail or the telephone. People may be “asking” things of search engines they might not even ask their colleagues, partners or best friends. The laws that regulate search surveillance and access to search engine logs by third parties have yet to adapt to this reality.

Ideally, search engine law and policy would not only minimize the collection of user data by search engines, but also place special restrictions on the access to search engine logs and guarantee a minimum level of transparency with regard to access to search logs by third parties. Without such restrictions, government agencies and other third parties will continue to line up to get access to these incredibly rich collections of data about the interests, intentions and online navigation of Internet users. In the meantime, Google regularly complies with valid legal requests for user data and gives no information about them as a matter of company policy.³⁹

Conclusion

Search engines are amongst the most used services by Internet users and are central to the navigation of the Web. In fact, they are an indispensable component of the Internet’s promise of democratizing access to information. Search engine law and policy should react by making freedom of expression a dominant concern underlying legal and policy choices with regard to Web search engines. Unfortunately, legislatures across Europe are lagging behind in facilitating the search engine environment in this manner. In particular, they have failed to provide the legal framework for search engines to develop and provide their services in a way that is consistent with the right to freedom of expression. The effective ability of users to find valuable information online without undue interference, the ability of online sources to reach an audience, and the ability to provide search services, to locate, select and evaluate information, are restricted in manners that are suboptimal from the perspective of freedom of expression. Therefore, in many ways, the current levels of access to information exist despite the law and not because of it. If one accepts that freedom of expression should be one of the dominant concerns underlying any law or policy for the Internet,⁴⁰ this situation calls for a change.