Why mobile data experiments are inevitable and good for society | Technology


"The importance of the data revolution is like electricity or printing," says Alex Sandy Pentland, co-founder of MIT Media Lab and one of the world's seven leading data scientists, according to Forbes. "It has completely changed the world, more than any technology of the twentieth century," he adds.

The controversial INE experiment that began on Monday in Spain is something that dozens of companies do without telling. The analysis of millions of phones to know how consumers move or to know the browsing history helps them to find out behaviors. It is already key to making decisions in the private sector. Why should citizens and public agencies not use it with all the precautions? And by the way why not take the opportunity to ask for responsibilities on who and how you use that information?

"The ability to collect data from everything is the most important thing that has happened since the 19th century"

The treatment of all these new data for the common good has been the battle of Pentland for a decade. Pentland talks with EL PAÍS precisely on a recent fleeting visit to Madrid to explain at the Club de Madrid how to handle data well.

"From a scientific point of view, the ability to collect data from everything is the most important thing that has happened since the nineteenth century: more than airplanes, cars or the Internet. We have data millisecond to millisecond of almost every human on earth," Pentland explains. We have not seen almost anything of what can be done with that. "It can be scary," says Pentland. But there are more options.

It will be perhaps the great debate of our era: "We have to integrate the data into the social contract," says Pentland. "We know how to handle money, how to manage the land, but how we will handle the data is a huge challenge," he adds.

Now the fear of losing something very valuable – our privacy – dominates the story. "It's something to think about. We have notions about the data recorded in our heads so deep that we don't even know they are there. And they are wrong," says Pentland. A November 15 report by Pew Research in the United States revealed that 81% of the public believe that the risks of data collection by companies outweigh its advantages and 66% say the same about the Government. The figures reflect a growing but misunderstood concern.

Pentland's work was partly at the origin of the EU General Data Protection Regulation, but today it goes further and proposes a collective solution: data unions. "It's like 120 years ago, when there were a few companies that dominated our work. What did we do? We formed groups to confront us. Now we need data unions. I don't think the government should be the solution to everything. The collective can solve it." Pentland explains.

Another proposal of Pentland to avoid misuse or loss of data is not to move them: "Do not gather all your data on a site, which is what we as humans tend to do, but leave them where they are collected and send pre-agreed questions," he explains. The algorithms that analyze this data should be open and revisable: "If you are a teleco, I am going to send you algorithm 23, pass it through your data and hang up the answer in open," he explains. Citizen control and transparency should distinguish hidden business from public good data.

The importance of diversity in your neighborhood

Why can our data be so important for the common good? Because they can save millions of lives or improve segregation in our cities. Pentland has its examples. Our behavior, for example, allows us to predict when we will be sick before we know it. Humans are beings of habits and 90% of our activity can be foreseen: you get up, go to work, come back. Then there is the unpredictable part: exploration, look for new things. "It's the most important part of your life, your curiosity," says Pentland. "The data shows that the more exploration, the better it will be in your life. When you see someone's exploration goes out, it looks like they will have problems. We are able to predict that," he explains. Now, a hospital can use that data, according to Pentland. "If you are a patient of the heart, they ask you to contribute your data. And if they see that your habits change, they call you to see if you are well," he explains. It is a process that the patient must accept, it is not involuntary surveillance.

"Diversity serves to predict growth. And when it falls, it is a crime predictor."

This is an individual example. But the vast majority of useful data collected is aggregated. "It doesn't take so much granularity. For social functions such as mobility or segregation, aggregate data is enough," says Pentland. In his research, he has worked with governments that have analyzed data similar to those that the INE will collect.

In Beijing they combined data from neighborhoods to find patterns of economic growth: "We saw that there are characteristics of diversity in neighborhoods that have more options to grow, it is a very strong relationship. Diversity serves to predict growth. And when it falls, it is a predictor of crime, "explains Pentland. In London they saw something similar: when visitors stop going to a neighborhood, it is a predictor that after two months the crime will rise.

With the Spanish Esteban Moro, Pentland has measured segregation in 10 North American cities with mobile data. They have looked at what places people go during the day. Each user has been assigned an income according to the place where he spent the nights. Therefore, each point is anonymous, but has an income assigned for its neighborhood. What have they found? "Half of the segregation in the US is due to your daily choices. When you go to eat, you enter a place that has people like you and not people other than you," says Pentland. The obligation of a city council is to look for spaces where citizens of disparate incomes meet. In Riyadh (Saudi Arabia), they found that the majority of those affected by chronic unemployment lived in areas where public transport was bad.

They are just examples of what will be in the near future: every good government will have its data lab to know its actions. It is not something that will excite politicians, who will face their prejudices with inexcusable results: "We must see it as an enriched census. Is it true that no one imagines running a government without a census? Now let's enrich it," says Pentland.

EL PAÍS published this summer summer data of Spaniards from the information of the mobile towers that a private company had. The INE seems to have been inspired to obtain similar and somewhat more refined data than by zip code. But the INE will still not know where each Spanish goes to work or on vacation, he will know where the majority of each group goes according to their place of origin. From the analysis of these data, ideas that have been difficult to demonstrate can emerge.

Politicians take advantage that we don't think

One of Pentland's goals is to find out how humans think and make decisions. We do many things without thinking, that's why we are so predictable. "Knowing how we mix our impetuous part and our reflective part is key. We see ourselves as independent and free will beings and such. But most of our life is not," Pentland explains. It is the distinction between system 1 and system 2, discovered by Israelis Daniel Kahneman and Amos Tversky. "Financial panics come from system 1, for example. It's not rational, you don't think about those moments," says Pentland.

That system 1 has a problem: it is automatic, visceral. "The facts and information do not serve to counteract it. So you have to find another way to communicate with people and probably to do politics." We don't have, according to Pentland, any explanation of how collective intelligence works: how humans make decisions together, for example precisely about politics. Pentland gives a wonderful example: "If you believe in the individual rational model, you should think that direct democracy, that everyone votes without legislators, would be the best. But all the evidence shows that it is the best way to provoke madness," he says. . So today we have representative democracy, which slows madness but leads to corruption. "There is still a better solution than having powerful and greedy people acting as representatives, but I don't know what it is," Pentland concludes.

. (tagsToTranslate) experiment (t) mobile data (t) (t) be (t) inevitable (t) good (t) society (t) alex pentland (t) co-founder (t) mit media lab (t) defend (t ) need (t) know (t) better (t) habit (t) technology


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.