Freedom, authority and knowledge on line: the dictatorship of the algorithm

• Author: Gianluigi Fioriglio
• Position: University of Rome 'Sapienza'
• Pages: 395-410

Revista inteRnacional de Pensamiento Político - i ÉPoca - vol. 10 - 2015 - [395-410] - issn 1885-589X

395

FREEDOM, AUTHORITY AND

KNOWLEDGE ON LINE: THE

DICTATORSHIP OF THE ALGORITHM

1

Gianluigi Fioriglio

University of Rome “Sapienza”

gianluigi.origlio@uniroma1.it

Recibido: octubre de 2015

Aceptado: noviembre de 2015

Keywords: Search engine, search neutrality, web neutrality, search algorithm, liability.

Abstract: Web search engines are a very important mean to ght infor-

mation overload’s consequences, but this makes them the gate to digital

information of any type and purpose. The cyberspace is accessed through

such complex and automated tools: software agents execute secret and

complex algorithms and make information easily reachable or hidden, but

anomalies and bugs may have serious consequences. Private entities provi-

de such services, on a global (e.g. Google), or local but relevant scale (e.g.

Baidu). After the Google Spain case, literature on the right to be forgotten is

growing. This paper aims at go further, investigating both the right to access

on line information and to be correctly and neutrally indexed by web search

engines. The law must regulate this topic, making them work in a neutral

and non-discriminatory way, even if they work cross-borders and are private

subjects. Otherwise, the Information Society will even more controlled by

the dictatorship of the algorithm.

1. Introduction

Legal literature on web search engines2 is growing, especially after a landmark decision

by the European Court of Justice: the judgement in “Google Spain SL and Google Inc.

v Agencia Española de Protección de Datos (AEPD) and Mario Costeja González” (13

May 2014, “Google Spain case”; European Court of Justice 2014). It held that the

right to be forgotten applies to search engines; in particular, in the European Union it

is possible to ask for delisting of web site links from the SERP (Search Engine Results

Page). Thus, even if a web page is on line and cannot be legally shut down, it may not

be indexed if certain conditions are met. The ECJ judgement is a landmark decision

because it makes a division among what can be easily found on the web, on the one

1. This is the revised version of the paper presented at XXVII World Congress of the IVR - International

Association for the Philosophy of Law and Social Philosophy (Special Workshop on “From Net

Neutrality to Net Protability? Law, Politics and the Internet”), Washington DC, July 2015.

2. This paper uses the expression “search engine” to indicate automated web search engines.

Revista inteRnacional de Pensamiento Político - i ÉPoca - vol. 10 - 2015 - [395-410] - issn 1885-589X

396

hand, and what can hardly be found, on

the other hand. A web content may be pu-

blished, but it does not necessarily imply

that everyone may rapidly nd it, answe-

ring to a reasonable expectation of privacy.

This aspect is crucial in the Information

Society, because search engines are the

primary gatekeepers to digital information.

Another aspect is perhaps even more im-

portant, but it needs deeper investigation:

is there a legal obligation related to web

site indexing? In other words, is there a

right to be indexed?

In fact, if we take into account only what

happens when a search engine responds

to a query (as in the case of the right to be

forgotten), we forget what happens before:

how does a search engine build its SERP?

How does it decide what pages can be

shown and their ranking?

This is not only an antitrust matter, because

it relates to all subjects whose content is

published on the web and indexed auto-

matically by search engines. “In an envi-

ronment where consumers are no longer

passive receivers of information, but in-

creasingly active contributors to the infor-

mation ecosystem, access also concerns

the (controversial) debate about the entit-

lement of users (as creators) to be integra-

ted into search indexes and ranking lists,

or at least the possible remedies against

discrimination in the indexing or ranking

processes” (Gasser 2006: 232).

Today, a search engine can decide what

can be accessed on the web; it can hide

and/or rank information making it hardly

be found (“By controlling the communi-

cation infrastructure of the Internet, they

have become information gatekeepers”;

Laidlaw 2008: 1143). It is fair: everybody

3. Moreover, “Users have become dependent on

search engines, viewing them as authoritative and

uses a search engine to nd something.

Nevertheless, it is also fair to ask for search

(or web) neutrality.

Before investigating this aspect, it is

useful to make some preliminary observa-

tions also from the methodological point

of view in order to clarify the perspective

of this paper. Search engines are the key

information retrieval systems of the Infor-

mation Society and this eld is not new

for legal informatics4. By adopting such

approach, law can be a meta-technology,

as argued by Ugo Pagallo with reference

to the “laws of robots”. In particular, he

proposes “to approach the laws of the law

establishing the conditions of legitimacy

for the design, production, and use of

robots, conceiving the law as meta-tech-

nology, i.e., as a means to govern other

technological means” (Pagallo 2013:

10). Thus, “once such tecnique regula-

tes other techniques and, moreover, the

process of technological innovation, we

may accordingly conceive the law as a

meta-technology” (Pagallo 2013: 11).

With particular reference to search

engines, it has to be considered that the

cyberspace actually is a complex digital

maze in which it is possible to retrieve

any information stored. As the library of

Babel, it may contain the solution to many

problems (Amato Mangiameli 2000: 36).

A search engine plays a key role: it is not

just an information retrieval system, but

reliable. Search engines have become the tools

through which the democratic potential of the

Internet can be advanced or hindered” (Laidlaw

2008: 145).

4. Legal informatics always studied information

retrieval systems and their legal issues.

Technological advancements led to growing

studies in the eld of articial intelligence in the

legal eld (including, but not limited to, legal

reasoning and automated application of the law).