Artificial intelligence in companies, here's when it pays to use it

An article by Francesco Berlucchi

In Plato's Phaedrus , Theuth, an Egyptian deity, presents his latest invention, writing, to the king of Egypt, advising him to spread it among his people because "it will make the Egyptians wiser and more able to remember." The king replies that the discovery of writing will have a disastrous effect, it will cause "forgetfulness in the souls of those who learn it" because, Plato writes, by trusting writing, people will "get used to remembering" only through it. Marco Passarotti, Director of the Interdisciplinary Research Centre for the Computerisation of Signs of Expression (Circse) and coordinator of the graduate degree programme in  Linguistic computing, chooses the Myth of Theuth to explain to the many students who participated in the event "How analyzing natural language with AI can improve decisions in the company" that, after all, "the fear of automation has always been there". 

"When there is a curve, you either go straight or you curve," explains Passarotti, during the second appointment of the series "Humane intelligence. Turning Artificial Intelligence into a positive technology". The meeting was promoted by the Humane Technology Lab (HTLab) of Università Cattolica, the Laboratory directed by Giuseppe Riva that explores the relationship between human experience and technology, and was moderated by Manuela Perrone, journalist of Il Sole24Ore (read the article on ilsole24ore.com). "Artificial intelligence has learned what it knows based on the data with which it has been trained," Passarotti continues. And he asks: "If artificial intelligence is based on an experience quantitatively unattainable by a human being, and if the hallmark of a good business leader is precisely experience, what is the role of the business manager? And what is the discriminating human factor with respect to artificial intelligence?"

The answer, according to Passarotti, lies in the fact that artificial intelligence is endowed with "knowledge", understood as the set of "significant correlations between information". This is the "great novelty" of the new technology which, unlike its predecessors, is not limited to the possession of "information" but goes beyond. It is "wisdom," on the contrary, that remains a peculiar characteristic of man. Who is able to "understand knowledge", to "know how to apply it", to "know how to do the right thing at the right time". Then there is "creativity": a good business manager is "one who knows how to make creative decisions," explains Passarotti. "Artificial intelligence is also partly creative, because it sees correlations between the data that we don't." Yet, according to the director of Circse, American linguist Emily Bender is right when she calls ChatGPT "a stochastic parrot, a machine that analogically repeats what it has learned from data." Because "the correlations that the machine sees it finds in the data, it doesn't know how to build new and creative correlations."

Data, after all, is the new oil. Giuseppe Riva, director of HTLab, reminds us of this with his usual concreteness , citing research published in the scientific journal Proceedings of the National Academy of Sciences (PNAS) according to which "it is enough to know one hundred likes, inserted on a social network by a person, to have a greater ability to predict his psychology than a partner could have". The theme explored in the context of "Electoral campaigns in the era of Artificial Intelligence" resurfaces, during the first event of this cycle of meetings. Yet, Riva continues , "if artificial intelligence possesses knowledge, wisdom lies in understanding. Knowledge is the ability to answer all questions. Understanding lies in the ability to ask the right questions, and to find the answers to those questions. This is what makes man wise, and it is a capacity still exclusively of man."

"All this data, which we often talk about, has so far been used by artificial intelligence to make predictions," explains Massimo Chiriatti, Chief Technology and Innovation Officer for Lenovo's Infrastructure Solutions Group in Italy and Professor of Artificial Intelligence and Natural Language Processing for Decision Making in the degree course in Linguistic Computing. "Today, the machine generates textual content, video, audio, computer code. Imagine a pen that is not passive but continues to generate content on its own. It does so on the basis of what we have done in the past." For this reason, the ideal condition of use, in work contexts, arises "when a company has a lot of data and tries to use it in the best possible way". The machine "doesn't decide, it can't tie cause and effect well," and above all "it doesn't have an intention." So there's no need to be afraid.

"The jobs that are disappearing are the repetitive ones," Chiriatti continues . But the prototypical example, perhaps, is translations. "In the past, we used to write programs with a lot of rules to translate, but in which we introduced exception after exception. At a certain point, the paradigm changed: machines were given many texts, not rules. In this way, based on the statistical connection, the translations are much better than they were a decade ago." And that same technology has not supplanted the human being, it has flanked him. "Today, translators who use AI can improve their income, because they produce faster and with more quality." In short, says Chiriatti, "technology helps us to study more, to improve ourselves, delegating the most repetitive jobs to machines."

"We are afraid that artificial intelligence will replace us, because we tend to humanize it," concludes Passarotti. "For the first time in history there is a dissociation between language and thought. We are used to the fact that those who interact with us are human beings. The challenge today is to stop humanizing artificial intelligence, and become more human, wiser, and more creative. So as not to be subjected to technology and, on the contrary, to use it in the best possible way in the business environment as well." The risk will be the widening of the "gap between those who do not know artificial intelligence, and will suffer it, and those who know how to exploit it". Because he has matured wisdom, "the understanding of knowledge."

The interview is published on Secondo Tempo.