A few weeks ago Malcom Gladwell published an interesting post in which he talks about his experience inside an automatic driving car, the Waymo, a project initially developed by Google. The brilliant Canadian sociologist tries to describe the ambiguous relationship between man and machine when it comes to driving (perhaps it would be better to say “being driven”) and concludes by saying that we will trust these cars to such an extent as to transform us from drivers reckless in pedestrians reckless, devoid of any fear of these cars. The theme, therefore, would seem to be that of trust: can we really get on these cars without the fear of crashing into an obstacle that is not perfectly identified? And, more generally, what degree of trust do we place in software that increasingly makes fundamental decisions regarding our life? A few examples: diagnostics, our job potential, customer care, the calculation of our economic reliability and even love relationships! On the other hand, the history of artificial intelligence is counterpointed with trial and error, outbursts of enthusiasm and cold showers.
From confidential computing to federated learning: new frontiers of artificial intelligence
by Andrea Frollà
The first definition of artificial intelligence dates back to 1956 and is attributed to Minsky, McCarthy, Shannon, Newell and Simon in a by now historic congress at Dartmouth College: it was stated that a machine is intelligent if it can do things that, in the same situation, they would do. the human beings. Furthermore, one of the first experiments of this type dates back to the 1960s: Eliza, this is the name given to the (perhaps) first chatterbot able to simulate a psychoanalytic interview in a very rudimentary way. Think of the expert systems that went crazy (also sociologically and philosophically) at the end of the 70s up to the victory of the computer in 1996 Deep Blue against world chess champion Kasparov. There was no doubt: these systems worked. But even then a short circuit occurred: although the statistics and the results were on the side of the expert or intelligent machine, the technologies did not always have the success that the data had hoped for: the goddess was missing. Trust, above all because those expert systems worked in extremely limited fields and for limited problems. Even today, more than half a century later, statistical research by the Pew Research Center in 2017 points out that the majority of Americans interviewed have doubts and concerns about the development of self-driving car technology (and other types of AI applications). ). Although the research shows some correlation between a lower level of education and a greater degree of concern about self-driving cars, in general the Pew Research Center report highlights not only a weak propensity to use these cars, but also a widespread concern about the increasingly massive presence of artificial intelligence in decision-making processes that affect our life.
The exhibition
If the epitaphs are written by artificial intelligence (in a deconsecrated church)
by Emanuele Capone
Since 2017, however, a long way has been made and to look at the matter with the eyes of rationality, there are very few doubts: these cars will learn to drive much better than us, they will have very accurate sensors that will allow them to increase safety in an unthinkable way. road, they will not be able to get drunk, they will not exceed the speed limits and, above all, they will not be able to get caught up in emotions.
Here I think that the crux of the matter lies precisely in this word: emotion. In fact, analyzing the argument using the frame, – or if we want the metaphor -, of trust we only grasp a partial aspect of the matter. In my opinion, however, the frame that we must use to understand this reluctance to use automatic guided machines is precisely that of emotion. In fact, the car – it is even obvious to say, just look at the commercials of the car manufacturers – is still considered, presented, “framed” as an extension of one’s emotional sphere: not as a simple means of locomotion, but as a means of transport. expression. It is perhaps in this sphere that we find it increasingly difficult to give space to artificial intelligence since it has taken possession of it in many other areas of our life as well.
The point is that technology continually promises us new spaces of freedom and greater security, and new ways of expression, but in a case like this, so practical and tangible, the balance between costs and opportunities of technology becomes much more difficult to weigh. .
From confidential computing to federated learning: new frontiers of artificial intelligence
by Andrea Frollà
.