When we do something online, we always leave a trace. We search for an address on Google Maps, a product on Amazon or a restaurant on Tripadvisor; we click “Like” on a Facebook contact post, re-share a tweet, publish a photo on Instagram and describe it with a hashtag; we read and respond to an e-mail, chat with friends and acquaintances on Whatsapp and Messenger, and sometimes we interact with business chatbot to obtain information that interests us. We do an hour running and use an app purchased on the Apple store that uses the GPS on our smartphone to know how fast we have run and how many calories we have consumed. If we carry out a search on Google while we are logged in with our account, the results will be different from public results because they are partly calibrated on what the search engine already knows about us.
The services we use record our buying behavior, our preferences and our social relationships, to try to know more precisely who we are. The aim is to accumulate information in large quantities, to be managed by systems able to efficiently intersect data that are gradually collected. The purpose is twofold: to know individuals and to have accurate information about larger social groups, also for predictive purposes. Naturally we are in the field of so-called Big Data: we have been talking about it for some years with reference to the dizzying increase in the volume of data and the speed of information flows on the net.
Having to deal with Big Data – and not simply with data – means dealing with large and complex information sets that require the definition of new tools and methodologies to be able to manage, extrapolate and process them quickly. What happens every day on Facebook is equivalent to billions of continuous scripts and rewrites on huge databases: all the likes, comments, reactions, preferences expressed, check-ins and geolocalizations, time spent on pages and video contents, tags and changes to contact networks are saved in real time and stored in the memories of thousands of computers scattered around the world. Most of the value that the company created by Mark Zuckerberg extracts from our actions is automatically managed by software: suggestions for new friendships or the advertising offered to us, for example, are managed by periodically updated algorithms. No human being would be able to manage the amount of information that is processed in real time by Facebook software, and then obtain useful data: for now, humans are left with control of notifications and political, ethical and moral controversies.
Why are Big Data not simply data, and why are our societies increasingly governed by algorithms
Further, without going into technicalities, what exactly are algorithms? To understand the issue better we can use a reference from the world of cinema: in a scene from the film The Social Network by David Fincher (2010), Mark Zuckerberg and Eduardo Saverin – who were later to become the co-founders of Facebook – discuss the possible adaptation of an algorithm created by Saverin. Zuckerberg’s idea is to apply an algorithm developed by his friend – designed to classify chess players – to another context: the classification of college girls. This is the birth of FaceMash, the predecessor of Facebook. The site is created in 2003, during Zuckerberg’s second year of college, to allow Harvard students to compare images of female students and indicate the most attractive from time to time. The FaceMash algorithm is nothing more than the set of rules followed by the software for proposing images to the user. From there, the step to the online and interactive version of the school yearbook with the names and faces of students was short: from FaceMash to TheFacebook, and finally to Facebook, Zuckerberg’s creation is something that today affects the lives of billions of people throughout the world.
So we know that an algorithm has to do with a set of rules. But what are these rules, and why do we use this word? The term algorithm has its roots in medieval Latin and refers to the name al-Khuwārizmī, given to the 9th century Persian mathematician Muḥammad ibn Mūsa because he was from Khwarizm, a region in Central Asia. In the Middle Ages, the derivative term algorismus was used to indicate the numerical calculation procedures based on the use of Indo-Arabic numerals. Today, the term algorithm is used to indicate systematic methods of calculation in general: specifically, uniform schemes and mathematical procedures for the resolution of a given class of problems. In somewhat more technical terms, an algorithm is an explicit and descriptive calculation procedure with a finite number of rules that leads to the result after a finite number of operations. Operations are nothing more than applications of the given rules which, in computer terms, are also called “instructions”. The computer algorithm is thus a set of instructions that must be applied in order to carry out processing or solve a problem. Algorithms are not to be confused with software, because the latter may contain internal algorithms but also contain non-algorithmic elements.
What role do algorithms play today?
At this point, one can ask how systematic methods of calculating and processing specific instructions have become increasingly important elements for our economy, information, culture and knowledge, health, education and for almost every aspect of our life in common. We are facing transformations that have long been ignored by educational and academic institutions, sometimes far from an effective understanding and management of the processes of progressive digitization of knowledge and of our social relations. The consequences of these “inadvertent revolutions” can also be found in the distance between generations, in the complications related to parent-child relationships and in family education, increasingly characterized also by technological mediation. The educational problem, in particular, is central to fully understanding what it means to completely entrust our lives to software and algorithms. Are schools and universities able to respond to the challenges posed by networked society and digital knowledge infrastructures?