Democracy and invisible codes E-Book

Sergio Amadeu da Silveira

• Herausgeber: Edições Sesc SP
• Kategorie: Geisteswissenschaft
• Serie: Coleção Democracia Digital
• Sprache: Englisch

Seitenzahl: 136

Veröffentlichungsjahr: 2019

Beschreibung

The series features in this fifth title the work of the professor and sociologist Sergio Amadeu da Silveira on the social implications of the technological development of algorithms. The author discusses the relationship between the advance of digital systems based on algorithms and the democratic debate. In an effort to understand how digital networks organize our daily lives, Amadeu looks into the role of algorithms in mediating and modulating public opinion. Citing key authors and practical examples, the book is organized into chapters such as "Democratic theory and the information society," "Freedom of speech and freedom of viewing," and "Can algorithms serve democracy?" The Digital Democracy series is published in Portuguese and English exclusively in digital format.

Leseprobe

I dedicate this book to my first grandson, to my daughters Bruna and Cecilia, to my sons Lucas and Pedro and to my partner Bianca Santana.

ACKNOWLEDGMENTS

I am grateful to my colleagues at the Laboratory of Free Technologies of the Federal University of ABC for opening new horizons. I would like to thank the staff at Sesc for their support and guidance. Also, I would not have been able to write this book without funding by Fapesp for the research program I coordinate, called “Algorithmic regulation in the public sector: theoretical and programmatic mapping”.

We accuse technical objects of turning man into a slave: that is perfectly true, but man is actually a slave to himself for he assents by giving in to technical objects; he surrenders to them like a soul is sold to the devil, for the sake of power, glory or wealth; temptation comes not from the object, but from what the subject thinks he sees in the object he mediates...

Gilbert Simondon

Models are opinions embedded in mathematics.

Cathy O'Neil

There is always a distinction between autocracy and democracy, since in the former state secrecy is a rule and in the latter it is an exception governed by laws that do not allow it an undue extension.

Norberto Bobbio

CONTENTS

Presentation

Introduction

Chapter 1

Algorithms and society

Chapter 2

Democratic theory and the information society

Chapter 3

How algorithms affect democracy

Necessary conditions and procedures for the existence of democracy

Algorithmic modulation of public opinion formation processes

Privacy, machinic servitude and pervasive surveillance by devices

Chapter 4

Algorithmic opacity, traffic concentration and economic power

The idea of democracy in the face of algorithms

The Invisible Algorithmic Factory

Freedom of speech and freedom of viewing

Chapter 5

Can algorithms serve democracy?

References

About the author

Credits

SINCE THE LATE 1990s, with the spread of the internet, words such as interaction, collaboration, exchange, recombination and sharing have come not only to organize the grammar of digital networks, but also to influence the social dynamic itself. It is a set of expressions related to forms of production and distribution of information and knowledge that uncover new scenarios, demanding reflexive efforts to understand its effects on communication and culture, as well as on education, economy and politics.

Today, the reach of the digital connection networks in a country of continental dimensions such as Brazil is evident. If, on the one hand, internet promotes an unprecedented enhancement in remote interactions and an exponential increase in access and production of content, on the other hand, there is a fierce dispute over attentions (and accessions) in its environment, which are increasingly narrowed to a limited range of platforms, sites and applications.

With the growing use of networks in the country, issues such as freedom, human rights, social equality, censorship, gender and race populate the daily life of virtual forums, often providing alternatives to the type of approach developed in traditional media, such as the radio, the TV and the written press. This is due, among other factors, to the relativization of the division between those who dictate and those who consume information, as this boundary is currently being erased.

Since the expansion of the network leverages, the multiplication of data volume and its corresponding dissemination in the public sphere, and it also stimulates the participation of an increasing number of people in the discussions on subject matters of common interest, we should ask ourselves about the real impact, in the public eye, of this way of circulating information and voices.

In this sense, it is promising to create a collection that aims to bring together Brazilian authors dedicated to thinking about the dynamics of digital connection networks, investigating their influence on the direction of democracy. Edited by the sociologist and PhD in Political Sciences Sergio Amadeu da Silveira, the Digital Democracy series invites researchers from the field of digital culture to scrutinize, from different approaches, the recent history of this ambivalent relation.

In this fifth book of the series, Democracy and Invisible Codes, the professor and sociologist Sergio Amadeu da Silveira addresses the relationship between the advance of digital systems based on algorithms and the democratic debate. In an effort to understand how digital networks are part of everyday life, the author looks into the role of algorithms in mediating and modulating public opinion. In his words: “Algorithmic structures, made up of databases, mathematical models and enabling software, have become a key factor in processes of public opinion formation and in the contest for voters’ political preferences”.

With a clear and direct language, the Digital Democracy series also seeks to awaken the interest of researchers in the area of technology and communication, as well as that of a more comprehensive readership, who are surrounded in their daily life by permanently connected technological devices. This publication in digital format makes use of a support capable to expand the possibilities of accessing studies about central aspects of contemporary life. In this way, it reinforces the role of reading as a key feature in education conceived on an emancipatory basis, using digital technology as a tool propitious for a critical, inventive and renewing social space.

Danilo Santos de Miranda

Regional Director of Sesc São Paulo

THIS BOOK DISCUSSES THE RELATIONSHIP between democracy and algorithmic systems, between politics and technology. Langdon Winner warned us that technologies can embody specific forms of power and authority, contain certain political properties.1 From that perspective, I will address here the link between the essential conditions for the existence of democracy and the increasing use of algorithmic structures that mediate and modulate our social relations.

Algorithms are never alone. They are part of an actant network. To understand them we must investigate their connections with the data structures that feed them and the systems that implement them. Algorithms are finite and logically chained routines that perform tasks based on received information. When we talk nowadays about big data, we are referring to technologies that use algorithms to manipulate large amounts of data.

Software is full of algorithms. The more computation, digital networks and cybernetic devices are used, the more we need to understand the algorithmic effects in the environments where they operate. We already know that they are not simple, much less neutral. Algorithms are performative and produce reactions, generating change in spaces and in those who do not always perceive their invisible presence.

These algorithmic structures, made up of databases, mathematical models and enabling software, have become a key factor in processes of public opinion formation and in the contest for voters’ political preferences. Social networking sites such as Facebook and Twitter, search engines such as Google and Yahoo, online programmatic advertising, ranking systems to define how institutions will treat us, cameras with face recognition software, among other examples, make use of such algorithmic structures and systems.

In addition, as the internet grew, so did the striking presence of algorithmic systems. Google was born in 1996 as a research project developed by Larry Page and Sergey Brin at Stanford University. They launched a search engine that sorted search results by page relevance based on the number of links to a page rather than merely on the number of times it was searched for. This automated decision was made by the algorithm called PageRank. Google soon became the most accessed node worldwide2.

The explosion of social networking sites further increased the importance of algorithmic systems. LinkedIn was founded in May 2003, followed by MySpace in August. January 2004 saw the birth of Orkut. Facebook and Flickr emerged in February 2004. YouTube was launched in February 2005, Twitter in July 2006. All these social networking sites had algorithms that organized the distribution of content published by users and of ads bought by advertising companies.

Facebook leaped ahead by creating the news feed in 2006. As reported by Eli Pariser, the news feed algorithm succeeded in gathering on each user’s home page the content created by his or her friends, becoming an actual customized newspaper. But, according to Facebook creator and owner, Mark Zuckerberg, by 2007 the amount of news published daily already exceeded the total produced by any other media outlet in their entire existence. The solution was to create a new algorithm called EdgeRank, which ranked interactions on the platform. The evolution of this algorithm depended on a constant feed of the data structure of each move made by Facebook users3.

The platforms turned into data devourers. With the advance of machine learning algorithms, the build-up of gigantic databases became indispensable for big online businesses. Google, Facebook, Amazon, Apple, among others, and the advertising networks started organizing mechanisms to capture personal data on unheard of scales. Offering what people were looking for meant knowing as much as possible about each person. That allowed platforms to offer their users things that could “improve their experience” and make them “more comfortable” in order to attract them through news and ads. Algorithms played a decisive role in directing online attention to specific nodes, websites and gigantic platforms.

Governments and other state agencies have also made progress in the use of algorithmic structures. Activities previously performed by civil servants are being replaced by automated systems. This undoubtedly provides the state with greater agility, but what else happens when algorithms start largely to define public interest?

My reflection here is of a critical nature. I do not intend to present a list of ways in which the implementation of algorithmic systems can contribute to modernize public services. Nor will I offer examples of how digital technologies can encourage and ensure political and social deliberation and participation, such as smart public consultation solutions, public policy-making with the participation of segments of society, process-building for smart cities, websites with budget simulations on likely public policies, among other. My approach is to point out the risks of algorithmic structures.

In chapter 1, I discuss the relationship between algorithms and the information society. In the following chapter, I address important elements of the relationship between democratic theories and the information society. In chapter 3, I examine how algorithmic structures and systems can affect democracy, especially in public opinion formation. In chapter 4, I analyze the relationship between the opacity of algorithms, the concentration of traffic in digital networks and the expansion of economic power. In the final chapter, I look into the necessary conditions for algorithmic structures to serve democracies.

This book is an effort to understand the sociotechnical networks that structure our daily lives. It is a technical and political work that aims to analyze the dimension and mediation of cybernetic technologies in the organization of societies. It is an essay that unveils the processes of machinic servitude and social modulation fostered by capitalism from the use of big data, machine learning and advances in artificial intelligence marketing.

Can one conceive a sociology of algorithmic systems or a political theory in which sociotechnical systems are actants as relevant as collectives and party leaders? Do discourses on the objectivity and precision of algorithms represent a positivism invigorated by the overwhelming and astounding amount of data? Raising doubts rather than proposing solutions, this text aims to invite readers to reflect on the sociotechnical directions of daily life. I hope its reading may instigate new and deep reflections.

1 Langdon Winner, “Do artifacts have politics?”, in: David Preece; Ian McLoughlin; Patrick Dawson (eds.), Technology, Organizations and Innovation, v. 2, 2000.

2 John Battelle, “The Birth of Google”, Wired, 1 Aug. 2005, available at: <www.wired.com/2005/08/battelle/>, accessed on: 15 Jan. 2019.

3 Eli Pariser, The Filter Bubble: What the Internet Is Hiding from You, New York: Penguin, 2011.

ALGORITHMS HAVE BECOME WIDELY USED and distributed devices in our society. Where are they found? In software, calculators, robots, automotive vehicles, aircraft, smart traffic light systems, online search engines, social media, apps like Waze, in various machines and instruments. They have already grabbed the headlines, as in the scandal of Cambridge Analytica, which aimed to govern the political decision of Facebook users, or made their way into face recognition software of surveillance cameras. The increasing use of algorithms follows the intense digitization of our communication, our files and our symbolic expressions. It also expresses the widespread automation of our production activities. The advance of information and communication technologies has made the use of software indispensable.

Algorithms are not software. They were born well before. The term gained prominence with computing, but its mathematical origin dates back to 9th-century in Baghdad. Historians argue that the word “algorithm” comes from a treatise by the mathematician Abu Ja’far Muhammad Ibn Musa Al-Khwarizmi, who lived between 780 and 850 in Baghdad, having worked in the House of Wisdom, a kind of center for higher studies, in the reign of the caliph al-Mamum (813-833)1. The treatise written by Al-Khwarizmi addressed Hindu-Arabic numerals. The original manuscript was never found, and we owe our knowledge of it to the Latin translation titled Algoritmi. There we find the origin of the term “algorithm”, which may also have been influenced by the Greek word for number, arithmós2.

Algorithm is a method to solve a problem. It depends on unambiguous instructions, logically linked rules and initial information. Algorithms handle the input data that will be processed according to defined procedures and generate results expressed as other data or information. When we enter the address of a given destination in a route-planner app, in general the algorithm also scans the GPS information installed in our cell phone. Other data are entered, such as traffic congestion along each route. Then the algorithm starts processing the information, that is, it will probably use methodology borrowed from graph theory to calculate the best route to the destination of choice.

Thus, an algorithm can also be viewed as a sequence of well-defined steps for the abstract resolution of a problem. One might say that it is a set of finite, linked instructions in a formal language, executable at a given moment. Mathematicians claim that the algorithm can solve a class of problems. For example, the sums of 2 plus 3 and 8 plus 9 are problems of the same class, that is, the sum of two integers. It is important to emphasize that a certain class of problems can be solved by several types of algorithms.

The researcher Pedro Domingos warns that “an algorithm is not just any set of instructions: they have to be precise and unambiguous enough to be executed by a computer”3. Therefore, not everything is algorithmizable. There are problems that can be defined with extreme rigor but cannot be expressed in algorithms. They are undecidable problems.

Historically, the algorithm occupies the central position in computing science because of the way that it encapsulates the basic logic behind the Turing machine. Alan Turing’s concept of a machine that could be used to determine whether any particular problem is susceptible to being solved mechanically was a highly original interpretation of the aim of David Hilbert’s famous project of formally deciding whether or not any mathematical proposition can be proved true. The algorithm, which Turing understood as an effective process for solving a problem, is merely the set of instructions fed into the machine to solve that problem. Without the algorithm then, there would be no computing.4

An algorithm-operated society is a mathematized society. The Encyclopedia of Mathematics, an open source project run by Springer and the European Mathematical Society, contains elements in the entry on algorithms that are key to understanding its social implications:

The scientific importance of the concept of an algorithm has been recognized for a long time. Since the earliest times, people have looked for constructive methods to solve mathematical problems. Such efforts usually benefitted from the new availability of suitable symbolic notations. This process has made a significant contribution to scientific knowledge, especially so after it had become clear that some problems are inherently unsolvable (squaring the circle etc.).5

The algorithmization of actions and interactions suggests a deep quantification and mathematization of social relationships. In a sense, this trend can already be found in the promises of modernity that ultimately led to positivism, to the scientifization of life, to belief in the infallibility of the method that would unlock the laws of the universe. Zygmunt Bauman argued that modernity waged a war against ambivalence, against indeterminacy, in the quest for order, control and the defeat of chaos6.