"When you don't pay for a product, you are the product"

Since beginning my doctoral research it seemed interesting to me to work on the automated actions that we increasingly encounter when we browse the Net: an ad targeted at us, some content, a new friend or a professional contact... I wanted to study these digital actions based on computational processes and how that phenomenon affects people's behaviour.

Yes, the Facebook News Feed (wall) is one of the most important news environments in the world. In Brazil, according to several surveys, this social network is the biggest source of information there is, ahead of newspapers. Eighty-six per cent of Brazilians with access to the Internet have a profile open in Facebook, which is over 80 million users. In the uprising against the government in 2014, 68% of people surveyed said they followed the events on Facebook.

Completely. Until that moment, Facebook was very simple. We all had our page but we looked at those of others to see the changes they had made. With News Feed, these actions were automated and the changes users made in their profiles became visible to others. It was the first controversy Facebook was involved in, because private actions were exposed to the public and that meant an attack on users' privacy, which caused a great backlash that forced the network to change its privacy policy. Nevertheless, News Feed survived. Although now it seems quite natural to us, at that time it was a very important change that affected how we use the Internet. From that moment, Facebook began to improve this mechanism and make it more sophisticated, with the aim of making it smarter and more personal to give users content more suited to their tastes.

It has very good strategies to obtain data. For example, all the questions it asks us when we register, which are increasingly more complex. It observes how long we spend looking at specific contents to determine what kind of content most interests us, and if we prefer pictures, texts or videos. It has what we call a centripetal force strategy, aimed at the centre. Notice the link that many pages now have with Facebook, which gives the network a great deal of new data, because we use other websites via Facebook. Another strategy is to try to prevent users from even leaving Facebook, so they consume most of their content there. This occurs on the mobile phone.

The concept of algorithm is over 2,000 years old, but now it is especially important for many reasons. An algorithm is like a recipe: it explains the steps you must follow to achieve a result. When we talk about the algorithm of Facebook or Google, we are not talking about only one algorithm, but a chain of algorithms that in their turn intersect with a chain of servers, with a chain of laws, of advertisers... It is a very big infrastructure that we label, mistakenly in my view, with the name of algorithm. Giving it this name is highly reductionist because in the end an algorithm is only a mathematical operation.

It is hard to say yes but imprecise to say no (laughs). The founder of Facebook, Mark Zuckerberg, says that his main objective is to understand human relations, the mathematical formula that makes someone interested in one type of content or another. It is something he talked about recently in a public meeting with Stephen Hawking. In my view, human patterns are too complex to be delimited in this way. The fact that someone clicks repeatedly on specific content does not necessarily mean that they are only interested in that content or that they want to consume more of that type of content.

Yes, it is a term coined by Eli Pariser in his book <i>Filter Bubble. What the Internet Is Hiding from You</i>. As he explains, the Net algorithm exploits data such as our search history, our location, or our past browsing behaviour to give us personalized results that, in some way, isolate us in a kind of bubble because it only provides us with what we would like, not what really exists on the Net. It is a very interesting discussion and highly disputed. This year Facebook has launched a study in which it has investigated a small percentage of its American users to conclude that it is the people themselves who cause the bubble, not the algorithm or algorithms. At this point I must also say that it is fundamental to have more information about personalization. Users must know that there are mechanisms to at least minimize it by using, for example, browser extensions.

Yes, and all companies allocate resources to this. Companies are now data driven: many products are designed with information based on data use. The more data there is, the better the products and the greater the commercial potential. There are people and institutions that are fighting to get this regulated. Europe, in this respect, has a fundamental role because of its critical position, non-existent in the United States. There are now studies that show that personalization has a strong impact on the diversity of information.

It is a company with specific interests that do not always coincide with those we have as people. Both Google and Facebook have a strong discourse in relation to doing good, making the world more open, connecting people and creating a more open Internet. I have been following Facebook for over ten years and in all this time it has talked about making the Internet more open when this is not what it is actually doing: it is closing it or taking it to Facebook. In South America or Africa, the Internet that is distributed free of charge is Facebook. It is important that people are aware of this, that they think about the fact that when you don't pay for a product, you are the product. But I want to make it clear that I like Facebook and find it useful.