SPECIAL COLLECTION OF THE CASE LAW ON FREEDOM OF EXPRESSION 2. Tension between freedom of expression and de-indexation Several experts, from various jurisdictions, have written extensively on the tensions that de-indexing represents to freedom of expression and the fragile justification the practice is based upon. The present section will give a brief account of some of the main arguments surrounding the discussion. The analysis will first discuss two foundational critiques and then explore the operational or practical problems posed by de-indexing. The first and most problematic aspect of de-indexing is that it directly constrains freedom of expression without complying with the basic test for imposing a legitimate limitation on said right. The right to freedom of expression is intended to protect a robust and uninhibited debate and thus also encompasses expressions that are intended to shock, disturb or offend. This has been the position of the European Court of Human Rights (“ECtHR”)3 and the Inter-American Court of Human Rights (“IACtHR”).4 In that vein, international human rights law has been consistent in developing a three-part test to justify a limitation to freedom of speech. The elements of such test, although they can slightly vary in different systems, are that (i) the limitation must be established by law, (ii) it must pursue a legitimate aim, and (iii) it must be necessary and proportional to the aim sought.5 De-indexing content can be hardly justified under this approach. Firstly, de-indexing is generally not based in a clear law drafted for that purpose. The application of de-indexing in the context of data protection schemes is sometimes justified, but often it is an overreaching application of remedies designed to allow a person to control how third parties use data that belong to her, not to allow her to shape publicly available records. The uncritical application of de-indexing as a remedy on data protection claims is problematic, for it fails to consider the freedom of expression interests—both at the individual and collective levels—involved, which are often not even considered by data protection legislation. In that sense, it is important to keep in mind that freedom of expression standards heavily protect information that has a public interest. Decisions that make it not as easily available should, thus, be considered a restriction on freedom of expression that should be interpreted narrowly and restrictively. Furthermore, the mere existence of publicly available information regarding a person does not entail an actual harm to that person’s reputation or privacy: this harm should be claimed and proved. The existence of a harm does not close the issue either: it should trigger a proportionality analysis in the face of a specific, carefully and narrowly crafted restriction on freedom of expression. Although in many cases de-indexing has been denied when involving information of public interest, a mere request to de-index certain content forces search engines to assess the public value of the information or the intrusion to privacy. Failure to do so expose them to liability, which produces the wrong kind of incentives on powerful intermediary actors in the flow of information on the Internet, and raises the risk of overreaching. 3 ECtHR, Fressoz and Roire v. France, HUDOC. App. No. 29183/95 (Jan. 21, 1999). 4 IACtHR, Eduardo Kimel vs. Argentina, ser. C 177 (May 2, 2008). 5 IACtHR, La colegiación obligatoria de periodistas, ser. A 5/85 (Nov. 13, 1985). 7