(versione italiana qua)
While in the previous post we focused on some characteristics of machine intelligence, in this one I concentrate on certain peculiar aspects of human intelligence that make it different from what I have suggested calling "mechanical intelligence" and therefore, I believe, impossible to achieve through artificial mechanisms.
These are elements that specifically and uniquely characterize the community of flesh-and-blood individuals, and in relation to which cognitive machines are intrinsically different from us.
A first difference stems from the fact that the information about the world that we acquire through our senses becomes representations, that is, data, for our processing, through the filter of our brain, which therefore makes us perceive a reality that is not necessarily objective, meaning universally shared. However, since people share the same human nature, despite their individual diversity, some degree of agreement can be reached about the objectivity of what is perceived, also thanks to the aid of language that allows us to explain and share descriptions of the sensations experienced. This does not happen between us and other animal species. All the more so, it does not apply to a cognitive machine that receives representations already defined by someone (based on criteria established by them but which may not be known and therefore difficult to share) or must construct them through physical acquisition sensors, which are also governed by other automata. It is difficult to argue that this allows for shared representations between humans and machines.
A second difference lies in the fact that, as already mentioned, cognitive machines do not have a physical body, and therefore cannot handle all those emotional aspects that intrinsically depend on it. The most recent advances in neuroscience tell us that when we are afraid or happy, these emotions are first and foremost physical reactions that occur involuntarily in our body due to perceptual events (but can also be triggered by the activation of a memory) and that serve the purpose, justified by evolution, of maintaining the body's equilibrium, for example by making us flee from a dangerous situation or remain in a positive condition. The role of emotions is fundamental in determining the birth and growth of social relationships and in evaluating situations. Their absence in cognitive machines is a difference that insurmountably places them in a separate class. Connected to this lack is the absence of self-awareness, that is, consciousness of oneself, a fundamental element for being able to "feel emotions." One cannot exclude, in theoretical terms, the possibility that, thanks to developments in artificial intelligence techniques, cognitive machines could be created that are somehow conscious of themselves. But even if this were to happen, and it seems very unlikely to me, it would still be the self-awareness of an alien species compared to the human race, because it would be based on different physical materials.
A third difference is that of creativity, intuition, of what in an underground, subconscious way makes it possible for us to "break out of patterns" and find decisive interpretive keys for apparently indecipherable scenarios. It derives, in absolutely unclear ways – given the current state of scientific investigation, from the interaction between the level of consciousness, namely the "place of the mind" where we have representations of the physical body, and the level of the physical body. Since cognitive machines have neither one nor the other, they seem to be clearly excluded from this possibility.
Certainly there are the very recent generative artificial intelligence systems that, starting from 2021, are capable of synthesizing, from a textual description, a photographic or artistic image, or even a video sequence, that realizes such description. Let's set aside the fact that these systems, not having a real understanding at a symbolic level of what they represent, can easily generate images or videos that are erroneous with respect to knowledge tacitly accepted and shared by humanity (e.g.: the fact that people have neither three arms nor three legs or that if two kittens are playing, a third kitten cannot appear out of nowhere between them). The central point is that the so-called creativity attributed to these systems actually resides in the formulation of what they are asked to create.
Everything else is nothing but the automation of a cognitive activity, extremely sophisticated and which I am in no way diminishing: it should be considered, however, that such automation relates to human intellectual capacities in the same way that an automated factory relates to human physical capacities. Clearly, the fact that all this happens at a level that, until a few decades ago, was reserved for the human species leaves us somewhat bewildered, but it is still a mechanical activity.
The fact that there are cognitive machines performing this type of work is certainly a positive aspect. In this sense I completely agree with the opinion of Charles W. Eliot, who was president of Harvard University for forty years, from 1869 to 1909, transforming it into one of the most important American universities: "A man should not be used for a task that can be performed by a machine."
[[The posts in this series are based on the Author's book (in Italian) La rivoluzione informatica: conoscenza, consapevolezza e potere nella società digitale, (= The Informatics Revolution: Knowledge, Awareness and Power in the Digital Society) to which readers are referred for further reading]].
--The original version (in italian) has been published by "Osservatorio sullo Stato digitale" (= Observatory on Digital State) of IRPA - Istituto di Ricerche sulla Pubblica Amministrazione (= Research Institute on Public Administration) on 19 February 2025.
Nessun commento:
Posta un commento
Sono pubblicati solo i commenti che rispettano le norme di legge, le regole della buona educazione e sono attinenti agli argomenti trattati: siamo aperti alla discussione, non alla polemica.