After having turned the dark web upside down in his previous adventure (“Il Giorno del Bianconiglio”, Chiarelettere, 2021), Leonardo Artico , the cybersecurity expert created by Alessandro Curioni , ends up again in the crosshairs of very dangerous opponents. In her second adventure, "Certe dead don't make noise" (Chiarelettere, 2022, pp. 300, also e-book) she has to deal with an ancient lover, in the meantime become the powerful manager of a big tech, next to whom a gray eminence materializes at the head of a “particular” nucleus of the secret services. Both have an offer for Leonardo Artico that cannot be refused.
Our protagonist - with his fellow adventurers the hacker Roberto Gelmi and the journalist Teresa Aprili - finds himself so involved in the very secret Da Vinci Project , which through the use of artificial intelligence aims to revolutionize the world of IT security.
To the author of the book, Alessandro Curioni, one of the leading international cybersecurity experts, we naturally ask if artificial intelligence is as "extraordinary" as it appears in his novel:
“If you think about it we call it Artificial Intelligence because we haven't been able to find a better term. After all, there is no unanimously accepted definition of what intelligence is. In any case, it is still a machine that manages to emulate some of the capabilities of the human brain: for example, image recognition. In fact, today we have the so-called weak, that is, specialized artificial intelligences at our disposal. Nothing to do with those represented in many science fiction works that are able to simultaneously simulate all the capabilities of a human mind. From my point of view, keep in mind that I deal with cybersecurity by trade: for me it is essential to know that where there are great opportunities, as in this case, there are great risks ".
So is artificial intelligence somehow already a reality?
"Absolutely yes. We often interact with intelligent algorithms even without realizing it. Think of all the virtual assistants, the self-driving cars starting to appear, the credit scoring systems. In short, we are surrounded by more or less intelligent objects ”.
Why does it appear as the technology of the future par excellence?
“I believe for the simple reason that he will be able to do something that is impossible for us human beings: manage and extract knowledge from databases of inconceivable size. What we define as 'big data' are inaccessible to us: we have a quantity of information available as never before in the history of humanity; yet, we fail to use them because there are too many. Many years ago the art critic Federico Zeri said: 'We are dying from excess of information'. Here, in this artificial intelligences can save us ”.
Yet, artificial intelligence is also dangerous ...
“It is for a few reasons. First, it is as fragile as it is powerful. By manipulating the databases on which she is trained, it is possible to induce her to develop cognitive biases. A famous case concerns the Compass algorithm used years ago in the United States to evaluate the recidivism of criminals. Eventually, it turned out he was racist. Secondly, there is the problem of 'dual use': what can be used to protect us can also be used to harm us. Algorithms have been developed that can emulate the writing style. This means that they may recognize a scam email with great accuracy, but they may also write a very credible one. Third factor is what is referred to as 'opacity'. By delegating decisions to artificial intelligences - think of a self-driving car rather than an algorithm that buys and sells stocks - it is very likely that we would not be able to understand its choices. This means losing the ability to control it ”.
Can artificial intelligence really put humanity at risk as we understand it today?
“I don't feel like excluding it. Some think that we are approaching the so-called 'technological singularity', which is the time when technical progress accelerates beyond our comprehension. The singularities have the characteristic that small variations can produce distortions of exceptional importance. I try to explain myself with an example. The environmental transition largely depends on digital transformation, the highest point of which is probably represented by smart cities. Smart cities whose management will largely be entrusted to artificial intelligences that will take care of monitoring and optimizing, for example, the consumption of buildings and traffic. It will be billions of sensors that will produce trillions of information then fed to thousands of artificial intelligences. We would find ourselves in the presence of a system of extraordinary complexity. If in 1975 Lorenz wondered 'can the flapping of a butterfly's wings in Brazil cause a hurricane in Texas?', I don't think the day will be far off when we ask ourselves: 'Can a failure of a traffic light in Singapore cause a blackout in London?'".