It is written AI, but it reads as informAtIcs

di Enrico Nardelli

(versione italiana qua)

In recent years, artificial intelligence (AI) has been on everyone’s lips. Even the average person has understood that it is something important, given its daily presence in the media, under the pretext that it "will radically change our future." While it’s true that all technologies have deeply transformed society, our human nature always remains the same. It’s worth asking ourselves whether certain messages are obsessively repeated for our own good or to serve other interests. Cui prodest? (= Who benefits?), one should ask, but that is not the subject of this particular reflection.

Back in November 2020, I was invited to a panel discussion on "Teaching Artificial Intelligence in Italy" organized by AIxIA (Italian Association for Artificial Intelligence) as part of their annual conference. During my talk, which I titled somewhat provocatively "It is written AI, but it reads as informAtIcs", I observed that the AI boom was finally making everyone realize the importance of informatics. Unfortunately, this awakening of attention was, and still is, accompanied by a view that considers AI as something separate from informatics.

In reality, AI is one of the oldest areas of informatics, because the desire to create something in our own image has always been strong in humanity. Historically, the term was coined in 1955 in a funding request that some USA informatics pioneers submitted to the National Science Foundation to conduct a two-month research project at Dartmouth (yes, just two months!). The proposal was based on the assumption that: «every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it». Although one always exaggerates a little when writing a research proposal to positively impress the funding decision-makers, the naivety of this statement is, in hindsight, somewhat amusing. This is especially true since the proposal went on to argue that «a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer». In its closing lines, the proposal offered a more cautious and appropriate description of the problem of artificial intelligence, describing it as the problem «of making a machine behave in ways that would be called intelligent if a human were so behaving». In effect, this similarity is precisely what we use to describe something non-human as intelligent, while remaining perfectly aware that it is an "as-if" and not "the real thing." This is a crucial aspect that we too often tend to forget when talking about technology.

In any case, the proposal was approved, and the meeting that officially introduced the term "artificial intelligence" to the world and launched the development of this field of informatics took place in 1956. Its products – also thanks to this name – perfectly illustrate, though not exclusively, the revolutionary power of informatics. It is the possibility of building what I have called "cognitive machines," meaning machines that are able to replicate the logical-rational cognitive abilities of human beings. Precisely because they are machines that do not tire or get distracted, they constitute a powerful aid for human cognitive activities. To put it simply, we can say that the automobile is to our legs what "cognitive machines" are to our brain.

However, as I wrote back in November 2018, together with an international group of experts, in a report on software source code as a world heritage for sustainable development, neither the general public nor policymakers are typically aware that these digital artifacts are fundamentally different from any human-made device of the past, which typically only enhances people's physical and sensory abilities. Computer programs (i.e., source code) are a special representation of human knowledge, not in the usual passive and static form used for centuries in books, but in an entirely new way. It is actionable knowledge, meaning knowledge ready to be put into action with the appropriate hardware and capable of interacting dynamically with the world. Source code therefore represents a mechanization of human knowledge that is unprecedented in the history of humanity.

These "knowledge machines" augment the cognitive capabilities of humankind, much as industrial machines have enhanced and extended people's physical and sensory abilities over the past three centuries. However, they lack true learning capability and the adaptability that characterizes human intelligence. Without a deep understanding of this fundamental difference between a cognitive machine and a person, the latter being the farthest thing from a machine that can exist, the role of informatics systems in society cannot be truly grasped. I discussed this in depth in my book, “La rivoluzione informatica: conoscenza, consapevolezza e potere nella società digitale” (= The Informatics Revolution: Knowledge, Awareness, and Power in the Digital Society).

With the arrival of the more powerful variant of artificial intelligence, the so-called Generative AI (GenAI), the situation risks becoming even more complicated. Our unchanging human nature invariably leads us to attribute more to cognitive machines than they are capable of, partly because of the words we use. The deplorable state of software development, which despite over 60 years of research is still incapable of producing artifacts with the same level of quality and predictability as other fields of engineering, does the rest. The consequences can range from the mere ineffectiveness of informatics systems — failing to perform the functions they were designed for or only doing so at the cost of great stress and enormous effort on our part — to the outright dehumanization of human beings, treated as mere things, disembodied objects to be disposed of at will.

Similar to the famous saying "beauty is in the eye of the beholder," we could say that for cognitive machines, "intelligence is in the brain of the reader". In previous articles, I have discussed several aspects that must be taken into account when considering the use of GenAI tools, from a societal perspective, for university education, and for school education, both in general and more specifically.

The terminological issue is therefore crucial: we shall return to it in more detail when we discuss the expression "artificial intelligence" in an upcoming article.

--
The italian version has been published by "StartMAG" on 18 february 2024.