As part of the digital festival “Futur en Seine” in Paris, a robotic hand sits on a wooden school desk and writes a punishment over and over again. A preventive punishment for a crime it has not yet committed. The phrase that was repeated on the paper in front of it read “I will not harm people”.
Don’t be scared! This is not some impending crime that is about to be committed against humanity. At least not as you imagine it and of course depending on which perspective you see it from. Artificial intelligence, then – which if you haven’t figured it out yet – is not a horror scenario. Some will think that artificial intelligence is something very distant from our everyday lives and have associated it with robots with red eyes or with science fiction scenarios. The truth, however, is that artificial intelligence is now in our lives every day. Typical examples are Siri, Apple’s digital assistant, or the well-known Netflix website, which is based on an algorithm that takes into account our preferences.
Before you get overwhelmed by the thought that humanity’s countdown is starting and you still haven’t learned it, let’s look at the matter a little differently. Why should it be either one or the other and not both together? Let’s start with the fact that there is a fundamental difference between the way computers learn and the way humans learn. A human, having an example as a criterion, can use it as a guideline and very quickly learn from it, resulting in imitating similar behaviors. On the contrary, a computer can only reach the same conclusions after being “fed” with thousands of examples.
So the question arises, are we ultimately “feeding” machines as much as we want them to “do” and maybe they will eventually not hinder human progress but rather accompany and improve their lives? One of the powerful figures in technology, Satya Nadella, CEO of Microsoft, has stated that “We are trying to teach machines to learn, so that they can do things that humans do today, but in turn they should also help humans by increasing their experiences.”
In case you are not convinced, do not forget that in 1997, Deep Blue, IBM’s supercomputer, was defeated by the famous Kasparov, in 2 chess matches. The third time it may have managed to turn the tables and the computer finally prevailed, but let us not forget that this happened after it was given the appropriate “food”, by a series of programmers who “plugged” the holes of errors or omissions of the original model.
Could artificial intelligence finally be as “threatening” as the human mind defines it to be?