Mind and Computer. 3 Essays on Human Intelligence and Artificial Intelligence

Since its foundation, in the second half of the 50s of the last century, artificial intelligence has been considered by many scholars to be a good model of human intelligence, therefore potentially able to successfully deal with all the activities carried out by man, even the most demanding on an intellectual level.

In the last 15 years, this belief, although not supported by concrete achievements, has been pushed even further, coming to theorize the next overcoming of human intelligence by artificial intelligence, which would materialize in the possibility of creating superintelligent systems, capable of intervene on their programs in order to self-improve, progressively increasing their performance, without any human intervention.

All this would inevitably lead to the danger that these systems develop a decision-making autonomy that puts them in a position to escape our control and pursue their own objectives, not excluding those of submitting us to their power or eliminating us en masse.

The essays collected in this booklet, though from different perspectives, aim to question the plausibility of these scenarios, starting with the fundamental assumption on which they rest, namely the one of the substantial analogy between artificial intelligence and human intelligence. The validity of this assumption is far from being demonstrated and, indeed, there is at least one element that leads us to think that it has no real foundation./p>

The intelligence of artificial systems is very fast – even millions of times faster than the one of humans beings – but it only does what it is programmed to do. It cannot go beyond the operating logic of the algorithms that govern it. It does not invent, it does not solve authentic problems, it is incapable of real discoveries or inventions. It can be said that it limits itself to re-proposing, although in variously elaborated forms and therefore difficult to recognize as such, what had been placed inside it. Not so intelligence of man, which shows itself capable, at least in some circumstances, of projecting itself beyond the information and regulatory criteria owned at a given moment.

This component of intelligence is usually referred to by the term creativity (or imagination), and can be defined as the ability to overcome the limits of the contingent situation by following new ways, which often lead to highly innovative results. Its existence is recognized by psychologists and by those who, for various reasons, deal with mind problems; even if no one seems to fully grasp its radical irreducibility both to the information processing processes typical of artificial systems, and, more generally, to the consolidated model of scientific explanation, based on tracing back to laws and theories of universal validity.

This lack of recognition – according to the author – would be the main responsible of the still widespread belief that artificial intelligence is entirely comparable to human intelligence and that it can one day exceed the performance of the latter.

These essays have already been published in Italian, on the web. They are now republished in English in Amazon's kindle format.

INDEX

About the author