(versione italiana qua)
Having completed with the previous post (Self-reproducing bits) a quick excursus on the most relevant fundamental concepts of informatics as a scientific discipline, in this second part of our stroll we begin a series of reflections on its impact on people and society.
Today's digital society is pervaded by cognitive machines, which carry out operations of a cognitive nature. I emphasize that with this term I am not attributing to these machines an intrinsic cognitive capacity similar to that of human beings, but only that the functions they mechanically perform are analogous to those purely logical-rational processes that people carry out. It is also worth adding that in individuals, since such processes occur in a mind embodied in a physical body, it is difficult, if not impossible in certain situations, to make them happen on an exclusively logical-rational plane.
This is a real revolution, the "informatics revolution," which I have characterized as the third "revolution of power relations," and which I discuss extensively in my book The Informatics Revolution: Knowledge, Awareness and Power in Digital Society, because for the first time in human history complex cognitive functions of a logical-rational type are being performed by machines. This third revolution "breaks" the power of human intelligence, creating artifacts that can mechanically replicate cognitive actions that were until now exclusive characteristics of humans.
In some ways these machines are reminiscent of those that, during the industrial revolution, made possible the transformation of society from agricultural to industrial: such industrial machines are amplifiers of human physical strength. In this case we have machines of a different nature and enormously more powerful: those of the informatics revolution are machines that amplify people's cognitive functions, namely devices that enhance the capabilities of that organ whose function constitutes the distinctive trait of human beings.
We have a technical revolution, namely faster data processing, but also a social revolution, namely the generation of new knowledge. The power that is being dismantled, in this case, is that of human intelligence. Humanity has always been, throughout its entire history, master and owner of its machines. For the first time this supremacy risks being called into question: we have machines that exhibit behaviors that, when enacted by human beings, are considered manifestations of intelligence.
We started with simple things, like sorting lists of names, but now we can recognize whether a fruit is ripe or whether fabric has defects, to cite a couple of examples made possible by that part of informatics that goes by the name of artificial intelligence. Certain cognitive activities are no longer the exclusive domain of humanity: we see this in various board games (checkers, chess, go, ...), once a unit of measurement for intelligence and in which computers now regularly beat world champions. We see it in a whole series of work activities, once the exclusive domain of people, in which so-called bots are now routinely used, computer systems based on machine learning techniques.
Finally, generative artificial intelligence systems have arrived (like ChatGPT, to be clear) that everyone has now heard about, which exhibit a competence in conversation with human beings that is objectively astounding. Unfortunately, this competence can be marred by errors or inaccuracies that are difficult to notice, unless we already know the correct answer. It happens that these systems exhibit excellent competence with words that describe the world but do not truly possess competence about the world, they do not have true understanding of the meaning of the words they use, despite appearances. In other words, we project onto what these systems produce the meaning that is within us: true intelligence is in the brain of the reader and not in generative artificial intelligence systems.
Postponing to future posts a more in-depth discussion of the role that such systems can play in the future, it is therefore interesting to touch on a couple of problems that apply to all cognitive machines, even to these more sophisticated versions.
The first is that in their current state cognitive machines have neither flexibility nor adaptability to change their way of operating when context conditions change. It is true that approaches based on machine learning techniques allow them to detect changes in their environment and adapt their actions, but this space for adaptation has severe limits. All possible future scenarios must have been somehow foreseen by the designers. People are intrinsically able to learn what they don't know, while cognitive machines can only learn what they were designed for. People have learned, through millions of years of evolution, to flexibly adapt to unexpected changes in the environment, while knowledge machines can – once again – only adapt to anticipated changes. We therefore cannot let them operate alone, unless they are in contexts where there is certainty that everything has been taken into consideration.
The second is that cognitive machines are completely detached from what it means to be human. Some see this as an advantage, for me it is an enormous flaw. I believe that there is no possibility of determining a single best way to make decisions. Those who think that through artificial intelligence human society can be governed in the best way for everyone are deluded (or have hidden interests). Since human society has existed, it has been politics' job to determine the synthesis between the conflicting needs that always exist in every community. And such synthesis cannot ignore our being human. The only intelligence that can make appropriate decisions in this context is the embodied intelligence of people, not the artificial intelligence of cognitive machines.
This does not imply that there is no role for cognitive machines. Their use should remain confined to that of powerful personal assistants, which relieve us from the burden of routine intellectual work, helping us not to make mistakes due to fatigue or oversights. But people must always have control and final decisions, especially those that – directly or indirectly – have significant consequences for other people, must always be made by human beings. We will discuss this again in future posts.
[[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 5 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.