Artificial Intelligence | 2 July, 2018

Artificial intelligence: what is and how we face it

Carlos F. de Córdoba Business Development

We speak of “Artificial Intelligence” (AI) when we refer to those machines or devices that are equipped with an intelligence similar to that of human beings. This term was used for the first time by the computer expert John McCarthy in 1956 and, despite the earliness of the announcement and the time that has passed, it is still not very common to find in our current environment.

Let’s start with a practical case:

This morning the car announced that it was short of petrol and taking data it knows (speed, location and consumption) and making rules of three, it concluded that it had enough to reach a nearby town. Immediately afterwards, and using GPS, it located on a map the four reachable petrol stations within its limited radius of action, although two of them were less recommendable due to the density of traffic and another one slightly deviated from the preset route. Impressive, but still far from being AI.

It’s easy to observe that the car had a lot of information that I didn’t know. It even determined that the nature of the announcement it transmitted forced people to look for petrol while trying to avoid random points like traffic jams, but it has not decided anything and it has not learned anything even after “seeing” my decision. Therefore, we can define AI as those algorithms that run on a computer platform and try to simulate the behaviour of the human brain.

It is no easy task, especially if we consider that we are tackling the body’s least known organ.

Expanding on and complicating the issue, human beings also have intuitions derived from repeated observations that are beyond the obvious. When we began the AI journey, ​​we had the feeling that with time we could make machines that could beat us at ludo, draughts or chess. Limited spaces and moves and some type of strategy to program; very simple in the case of draughts and very complex in the case of chess but both feasible. But what about poker? How do I program cunning or deceit and what is its strategic moment?

With this introduction, it seems clear that we need help. It’s time to find inspiration in nature; if we analyse any natural entity, we observe that intelligence is interrelated Moreover, interrelated and based on communication. To give an example that is as clear as it is unsettling: What is the organism, the ant or the anthill? Attempting to answer this question, scientists have concentrated on creating small cosmos of specialised artificial intelligence, abandoning, for the time being, the creation of a general AI.

If we achieve some degree of perfection with this cell-type strategy, the next step would be to emulate the neural networks making up our brain, use electrical impulses as a medium to exchange information between each other and condition the reactions of algorithms, not just what they know how to do for themselves, but according to what others “tell” them.

It seems that things, at least for the moment and with some obstacles, some of an ethical nature among them, which we will talk about later, follow these lines.