Algorithmic society: invisible intelligent agents

It is necessary to build new public policies for the algorithmic society.

29/01/2020
  • Español
  • English
  • Français
  • Deutsch
  • Português
  • Análisis
algoritmos_tecnologia.jpg
-A +A

It is very common to find journalistic articles that refer to algorithms. Whether in the economy, technology or lifestyle section, these mathematical processes have in a short time become a permanent part of our daily news.

 

As algorithms become the invisible "intelligent agents" in the redefinition of the digital world, it is urgent to talk about how to alert citizens to the impacts of data systems that make decisions about our opportunities, rights and vital possibilities. If one day you are denied a job, medical care, entry into the banking system, access to scholarships or work contracts because an algorithm dictates it, who can you turn to?

 

An algorithm, as well as the data with which it is trained, can be biased from its very conceptualization. Does the algorithmic society reproduce the concepts of the model of distribution of wealth and of inequality in access to knowledge?

 

Several authors, like Cathy O' Neil in her book "Weapons of Math Destruction", have denounced that the development of these models embedded in software are not regulated by States. Thus, they are creating a dual society in which the rich have the privilege of personalized, human and regulated attention while vulnerable groups are condemned to the results of "intelligent machines", in which there are no transparency, rights or clear procedures for appealing against algorithmic decisions.

 

Many of these issues deserve deep social reflection and building techno-political and social consensus on what is and what is not acceptable and desirable. But to begin to articulate this consensus, we need tools to help us understand what is happening. Can public visibility of the algorithms and training data be such a tool?

 

Redefining the digital world

 

Understanding the scope of the algorithms and their application contexts will help us understand how the digitalization of everyday life and our relationship with digital platforms and services is guided by a computing infrastructure that shapes the process of the algorithms. Therefore, it is necessary to understand, discuss and criticize the way algorithms, which are unknown to us, mark our days.

 

An algorithm is defined as "an ordered and finite set of operations that must be followed in order to solve a problem.” These actions are trained with data to obtain conclusions and knowledge.

 

The expansion of calculation methods expressed as algorithms, some available since 1950, is largely due to the encounter with application software (or apps), large-scale processing capabilities and the increasing bandwidth of the Internet, which today form the territory of Big Data.

 

Calculations penetrate so intimately into our lives that we cannot clearly perceive how they transfer our data to statistical infrastructures located on distant servers. Thus, a growing number of knowledge domains such as culture, knowledge and information, health, city, transport, work, finance and even love and sex are mediated by algorithms.

 

The first major expansion of Facebook coincided with the first economic crisis of 2007, which boosted the development of digital platform services aimed at individualizing, orienting and guiding each Internet surfer. By 2015, Facebook had 1.35 billion registered users, who communicated in 70 languages and used 50,000 servers (computers with high processing and storage capacities). Facebook has thus become one of the "owners of the Internet", controlling Instagram since 2012 and WhatsApp since 2014; a huge business representing more than US $25 billion per year and a volume of data far beyond the millions of dollars estimated.

 

The political value of the permanent data flow of this digital platform was visible in 2013, when the US National Security Agency recognized it was using Facebook to track citizens who "innocently" contributed their data, information that had been a target of intelligence work for years.

 

Something similar happens with search engines such as Google that carry invisible cookies (small software applications with algorithms) that make it possible to identify users and map out their web browsing and depth of penetration on each website, data that used to configure new digital business products.

 

Two dynamics are advancing which bring us into an "algorithmic society". The first is the digitalization of society with a growing demand from citizens to be included; the second is the development of processes.  The latter deliver to the computers, by means of software, the mathematical instructions to classify, order, group, predict, treat, aggregate and represent the information by means of increasingly unnoticed data. People's movements, shopping bills, clicks on the Internet, online consumption, time spent reading a digital book, time spent listening to music and watching video on demand, are all encrypted by algorithms, which classify and predict our present and future consumption.

 

Omnipresent in our lives, algorithms are presented as "mysterious", since we lack knowledge about their existence and functionality. A new religion with new acts of faith. We rarely question how these processes of logic and calculation occur and the worldview they entail.

 

Digital society has put us citizens in a position of digital consumers; we know how to operate devices and we consume content through apps. We do not know about the processes and procedures that regulate digital society. For example, that the algorithms used by the technological giants, known as GAFAM (Google, Amazon, Facebook, Apple and Microsoft), are not always fair in the decisions they make.

 

The algorithms operate inside "black boxes"; we do not know their functionalities. The value that moves the algorithmic calculation is none other than that of personalization, of individualism. This bias in the design of the software that implements the algorithm responds to political and cultural models of social relations and of distribution of knowledge and wealth. These biases can also affect decisions such as health insurance or prepaid medical care, schooling or criminal records, acceptance of a job application. The data only speak to the questions and interests of those who consult them.

 

In today's culture, algorithmic calculations capture desires for freedom and personal services; where individuals, through their representations, ambitions and projects, think of themselves as autonomous subjects, beyond inclusive or exclusive political models.

 

It is necessary to build new public policies for the algorithmic society. The processes and biases must be made explicit to citizens, as alerts in the Internet domain and in State Statistics. The State has to guarantee citizens the visibility of the algorithms and training data.

 

It is necessary to know the political and cultural vision that is implemented in the algorithmic processes and the bias of the data that give results as truths. A critical look at the functioning of hidden calculations must be established. It is necessary to know what meaning and goals the algorithms carry out. These meanings constitute a possible world where the recognition of merit is unimpeded; where authority is obtained only through quality and resilience.

 

The GAFAM aim to install an invisible technological environment that allows people to find their way around without contradicting them. A large part of our daily choices are made by a socio-technical infrastructure; buying a plane ticket, automatic language translation, finding the best restaurant, getting a personal appointment, filling the refrigerator or recharging the metro fare card.

 

With the GPS we have lost the landscape. Algorithms guide our preferences and bind our choices. They make the liberal dream of unfettered choice come true, but this dream also hides its other face: an algorithmically guided freedom.

(Translation ALAI)

 

- Alfredo Moreno, Professor of ICT at the National University of Moreno. ICT Engineer at ARSAT. Member of ticdata.com.ar  @ticdata2

 

https://www.alainet.org/fr/node/204468
S'abonner à America Latina en Movimiento - RSS