Few concepts have fascinated the human being on a par with Artificial Intelligence, fueling the imagination of science fiction authors of every age: the idea that man could artificially “create” an intelligence, or could even create demiurgically new consciences, has first seduced and then frightened hordes of scientists, philosophers, theologians and economists.
But just as there are few other topics related to information technology that have fascinated experts and non-experts, there are just as few other topics that have been equally misunderstood, mystified and deformed in their essence because of the need to write impressive sounding titles or take the credit for yet another proposal for an “innovative” law. And in a moment like the present one, in which there is more and more talk of Artificial Intelligence because applications related to this theme – wrongly (almost always) or rightly (very rarely) – are spreading with increasing frequency , it is essential to have clear what is meant.
The concept of Artificial Intelligence is actually very vast. In fact, this definition covers the most diverse phenomena: from the computers of 2001: A Space Odyssey and the robots of the Asimov sagas, more human than humans, to automatic call center responders which undermine the patience of even the most patient of humans. The first belong to the type of Artificial Intelligence defined as “strong”, that is, that which is said will even achieve self-awareness and which today (and for a long time to come) is relegated to science fiction movies. The second are a different matter. “Weak” Artificial Intelligence (which has nothing to do with Vattimo and Rovatti) is dedicated to analyzing or solving specific problems, sometimes computationally too complex to be analyzed by the human mind. This is the case, for example, of the Deep Blue computer which already managed to beat chess champion Kasparov some time ago but which, apart from playing chess, would not be able to do much more.
Weak Artificial Intelligence in the future?
It is weak Artificial Intelligence that we will not only have to deal with in the next few years, but which is already present in many of today’s applications. An AI that does not aim to “replicate” the human mind but, much more simply, perform specific functions better than human beings. Based on automatic learning, on automatic analysis of large amounts of data and algorithms sometimes able to change themselves to adapt to changing external conditions.
The applications are already endless today: they can be used to suggest “things people like” related to user behavior, for example on social networks, search engines or online stores; for crime prevention with the analysis of images captured by video circuits in cities or in public places; in the analysis of data aimed at forecasting customer purchases; in home automation; in the prevention of frauds which, from analysis of the customer’s habitual behavior, succeeds in intercepting possibly malicious exceptions; in video games; in the much feared and debated self-driving vehicles, and even in journalism, with automatic systems for writing news.
In this regard – for example – recent research has found that, in certain categories of articles (sports, finance, weather forecasting), a representative sample of US readers preferred articles written by a computer to those written by “human” journalists. But weak Artificial Intelligence is used today even for composing music by adopting the style of the great composers, and for developing commercials in a completely automatic way.
What is the challenge of AI?
In short, if we are far from self-aware computers and Matrix-style scenarios, weak Artificial Intelligence is only so in name, because the impacts on the economy, on the market and on society are and will be increasingly stronger. But we should not forget that all the problems we will have to face in the next few years will not depend on treacherous Matrix-style computers, but rather on rules that completely unaware computers will be forced to follow
The real challenge, therefore, is to understand how attention should be placed on who is writing the computer algorithms of tomorrow, and how these algorithms will be written. Because an increasingly important part of our lives will be regulated on the basis of these algorithms.