It is undeniable that our lives are surrounded by technology twenty-four seven. For instance, almost everyday, most of us access Internet, buy items online, use social medias (like Facebook, Instagram, YouTube or Twitter), read emails, etc. When accessing these pages, besides the content they offer, we are bombarded with suggestions of people to follow, news to read, pages to access, videos to watch, songs to listen to, things to buy, and so on. A curious one could ask: how do these websites know things we like? Are they watching us?

To answer this kind of question we need to dive into an interesting perspective: how technology underlies technology. Once we are capable of understanding this point, we will have the ability to answer so many more questions (and not only “why Facebook always shows the people I like most”).  For example, we will get the idea of how self-driving cars work, how Siri can talk to us or how a computer can defeat the world’s champion in a chess game.

In this way, to extract to proper knowledge from “how technology underlies technology”, we must be familiar with terms like artificial intelligence, machine learning and data science. I know, when we listen to people talking about this kind of content, our minds tend to wander into images of “super-intelligent robots” that will replace human beings in every single task. I am not here to say these “super-intelligent robots” will never exist, but let’s see a more realistic picture of what is happening nowadays.

The first concept mentioned is “artificial intelligence”. This term, defined many years ago, is, basically, the idea of computer systems being able to perform tasks that require human intelligence to be executed, such as visual perception, decision-making, speech recognition, etc. In order to develop these “intelligent systems” (or our fanciful “super-intelligent robots”), the programmer can take different paths, and, one of them (very common at the moment) is the use of “machine learning”. Machine learning brings techniques that leads the system to learn and improve from experience, without being explicitly programmed; to do so, the programmer shows to the computer many examples of a particular problem and the computer creates, by itself, the rules to return the correct answer when a similar experience is showed again. All the examples given to the computer during the learning process can come from different sources; currently, the most ordinary alternative is the use of databases which contain tons of required examples. Here is where the term “data science” appears: the programmer will need to work with raw data and, through the application of several techniques, will have to structure this data in such a way to make it meaningful and proper to be presented to the “intelligent system” about to be developed.

Based on these ideas, scientists are developing many and many intelligent systems, which are now having influence on our daily lives, even when we do not know they are being used. Many examples, besides those given at the beginning of this article, can be listed. Some weeks ago, a publication on the journal Nature Communications, showed an app developed to detect patients that present anemia just by analyzing pictures of their fingernails. A few months ago, another publication, also on Nature, used the same idea but now to classify skin cancer, which is said to be a complicated task due to the fine-grained variability in the appearance of skin lesions.

For many, it is strange to think that robots are the most important players, providing liquidity, in the most important financial markets around the world. They are there, untiring, all the time, making non-emotional decisions, faster than every single human being in this world. In the same way that people can buy a new self-driving car, they can also hire specialists to create new algorithmic trading systems for them, which are going to use the cutting-edge artificial intelligence techniques. How should we play in this new reality?

Undoubtedly, we have reached the point of no return. Although it may seem scary at some level, it is essential to experts from different fields to get used to this new way to solve the world’s problems. If we guide ourselves taking the past as example, it is possible to remember that, years ago, when the computers emerged, primarily those who developed the ability to use this strange new object were preferred to get the best job opportunities, because, instead of fighting against the natural evolution, they were smart enough to become adapted to it. Nowadays, how to use a computer is not a problem anymore. However, new technologies are erupting every single day, and that keeps our paradigm in the same level: the ability to understand, create and manage these “super-intelligent robots” in our own favor is the only chance we have to avoid being knocked down by them.

This article was produced by
Isabella Munhoz, M.Sc. candidate (USP)
Humberto Brandão, Ph.D. (UFMG) and Kaggle Master