What artificial intelligence teaches us about what intelligence really is


The arrival of artificial intelligence to the general public has meant the beginning of a debate on whether it is possible for this automatic system to replace human intelligence.

Multiple authors, from Noam Chomsky to epistemological philosophers , raise a series of doubts about this matter: Chomsky detects in her the banality of the evil of the philosopher Hannah Arendt, and others consider that she has no capacity for the abductive judgments of our mind, it is that is, for the generation of new logics in the unexpected progress of rationality.

Many authors think that this “intelligence” is far from being creative, it is not generative, although it is capable of producing apparently new texts. But really, what is the difference between artificial intelligence and human intelligence, just plain old?

Select, combine and compare

The authors who study compositional or synthetic intelligence, which is what can generate new ideas, have established that the human brain that processes information intelligently and creatively performs three different activities:

  1. Selective combination: joining and combining pieces of information in specific ways. It allows innovation by changing the order of the elements, and many inventions are the result of an innovative combination or association, such as the vision of the double nature of the electron.
  2. Selective comparison: projecting chains of associations onto new contexts, to observe the parallels, that is, making analogies and metaphorical projections. It allows establishing parallelisms and launching hypotheses from one field to another. For example, seeing the molecular structure as a spiral staircase or that of the nucleus of an atom as a small solar system, selective comparisons that generated crucial advances in knowledge.
  3. Selective coding: synthesizing, removing, and polishing structures until they lose ambiguity and superfluous information. This is the crucial principle in inventive and creative advances in all fields. To give a simple example: Rembrandt renounced the use of the blue color range in his paintings to enhance the expression with the other colors; o Frank Capra always recommended eliminating rolls of film to improve the quality of the whole.

Artificial intelligence uses all the information that we upload to the network, makes constant comparisons and combinations of elements, being able to track and present compositions on any theme or text. The resemblance to creative operations exists, but there is a radical difference. Which is it?

Shape, elegance and aesthetics

The French mathematician Henri Poincaré came to the conclusion that access to innovative knowledge was not a mechanical operation, nor did it come from rational logic. Acceding to a new idea, he said, was a “feeling of the form” of that innovation.

As he explains in his work Science and Method , when he discovered solutions in his search for mathematical advances, he experienced an impression of form, which was inexorably linked to the discovery, and on whose capture the subsequent development of the entire advance depended.

This experience could be considered aesthetic, and Poincaré came to experience the elegance of solving a mathematical problem. What did Poincare mean?

Jump to the future

Authors who study creative intelligence effectively register the paradox that when an innovation is created in any field, knowledge that did not exist before is “taken out of nothing”. This operation cannot be generated automatically or systematically. What there is in creative invention is a “leap into the future” through a new language, as Antoine de Saint Exupéry explained in his Cuadernos De él.

According to Saint Exupéry, the relationship between elements, which is previously unthinkable and improbable, is built through language. And this operation is a phenomenon that the creators see linked to the sudden vision of a new form, composition or structure. Once they are expressed and shared, they radically change our reality. But we cannot plan their appearance because they belong to the future that they create themselves.

The changing language

Only through a language capable of infinite combinations of finite elements, which can break its own rules and establish or shape new elements, it is possible for these jumps to occur.

According to Poincaré and Saint Exupéry, then, the invention is not something that results from the automatic compendium, nor from the summary of a profuse set of data, nor even from previous learning experience. It is about something different that is related to a “disruptive” use of languages ​​and codes.

When a creator, a researcher or a scientist innovates, they do not follow common usage patterns or compile universal information. Your selection of items can be completely inexplicable. They may even be unaware of basic information, and this precisely helps them to innovate. The analogies with which they work may escape all plausibility, but nevertheless, they will be like Galileo’s, essential. And his creative syntheses will surely be both new and amazingly appropriate. This is part of the advancement in knowledge, which cannot be automated or manufactured.

In a very recent investigation, precisely on the decline of scientific expansion, the significant disappearance of the use of disruptive languages ​​in all fields of human scientific progress is perceived.

Work with what is already known

Artificial intelligence cannot generate new information or use language to generate new thoughts never conceived by human beings, because its working base is what is “already known”, what already exists, information universally shared on the web.

It is an excellent medium for processing or disseminating what is already known, but not for thinking about the new. It is tightly tied to the past, and its structures and codes repeat and reiterate what already exists, what was previous.

These systems can go a long way in freeing reporters from uncreative tasks and may even force practitioners to do more than parrot conventional information from familiar sources.

If we learn to distinguish between artificial intelligence and creative intelligence, perhaps we can also learn with this that many of the forms of behavior that we consider intelligent today are nothing more than simple exercises in redundancy, superficiality, and even stupidity. Artificial intelligence can teach us this, at least.

Author Bio: Eva Aladro Vico is Professor of Information Theory at Complutense University of Madrid