artificial intelligence it started as something very specialized that only a few people could use. However, today anyone has access to it through algorithms such as, for example, ChatGPT. For this reason, many people warn about the importance of designing these algorithms properly, explaining what their goals are so that they don’t give us unpleasant surprises. One of those who paid the most attention to this issue is the philosopher Nick Bostromwho in 2014 explained with a very simple example what can happen if it is done incorrectly.
As an example, he cited artificial intelligence, whose task is to produce paper clips. No other goal is set for him, only this one. In the long run, according to this AI ethics philosopher, destroy humanity.
He used this example because it seems very innocent; but according to him can be extrapolated almost any artificial intelligence. For this reason, it is very important to define their goals and make it clear that these algorithms should serve humanity. If this is not done properly, they will have no reason to try to protect him.
How can an artificial intelligence that makes paper clips destroy humanity?
It is obvious that all this hypothesis that Bostrom raised to warn of the importance of goals. This artificial intelligence does not exist. But still, let’s see how he could destroy all people.
This would be a simple algorithm with the sole purpose of producing as many clips as possible. It would start doing this non-stop, but since this is an algorithm capable of calculating all the possibilities, there would come a point when I would take people as a nuisance. At any moment, someone could turn it off, and then it was impossible to shoot as many clips as possible. In addition, people have atoms that can be turned into more paperclips. They spend.
This is an extreme example. But it helps to see the importance define goals artificial intelligence algorithms as they become more efficient.
They must serve the people
Giving his example, Bostrom explained that one of the goals that needs to be defined is that artificial intelligence will always be must serve mankind. If it is not made clear and only works for the benefit of one person or group, the rest will be considered unnecessary.
In short, artificial intelligence is useful. A lot of. However, limits must be set and it must be very clear what it is for and how it is supposed to work. As Bostrom once said, “We must be careful what we wish for. overmindbecause we could get it.” In fact, it may even go much further than we want. To prevent this from happening, algorithms must be carefully thought out before they are developed. Clearly, now that it is starting to boom, there is much legislation to be passed and many ethical issues to be discussed. But beyond that, there is still a lot of technology that needs to be carefully considered so as not to surprise us.
Source: Hiper Textual
