According to experts, all artificial intelligence has security codes, barriers. This is what makes them, for example, not have homicidal intentions. Until then, everything is logical and reasonable. However, Former Google CEO Eric Schmidt is worried about it. In his opinion, these measures can be skipped, and it is not at all difficult to do so.
In other words: the technology expert, who for years was the head of one of the largest companies in the world, has joined those who view AI with suspicion. It does not at all rule out the appearance of what he calls «homicidal AI models». But what exactly are they? And even more importantly, how could they be a threat to people?
From Google to being afraid of AI
Something is happening with artificial intelligence. At first, no one gave much importance to the few experts who spoke that AI could become a danger to humanity. They looked like fanatics who had seen some science fiction movie along the way. Matrix o Terminator and they would have come up. Now, however, that is no longer entirely the case.
More than anything because they are not isolated cases. More and more specialists are sounding the alarm, either in one way or another. Some talk about extinction, others talk about massive unemployment, but everything always has a rather apocalyptic touch. Now, to this group we must add an eminence, Eric Schmidt, who for years was in charge of Google, no less.
During the Sifted Summit technology conference in London, Schmidt was asked about the dangers of AI. and he He had no qualms about even equating his threat to that of nuclear weapons.. In his opinion, there are security measures that all companies put in their models. The problem is that they don’t seem very reliable… at least in their opinion.
His words could not be more forceful: «There is evidence that you can take models, closed or open, and hack them to eliminate their security barriers. «A bad example would be if they learned to kill someone.». Through a malicious act, any hacker could access the orders that the AI has blocked. But they are there, despite everything. And that is a danger.
The risk of uncontrolled AI
In this way. Eric Schmidt joins the many experts calling for some form of regulation for AI. In these moments, There is no international regime that prevents this technology from expanding without any control.. So no, there is nothing that prevents artificial intelligence, sooner or later, from being used for what it calls «destructive purposes.»
A vision of AI that is not at all reassuring, which coincides quite a bit with that of other specialists such as Nobel Prize winner Geoffrey Hinton. Is it possible that this is not the paranoia of a few exaggerated people? If so, it is very likely that measures would have to be taken. But companies like OpenAI or Google itself seem more concerned about other things.
Get to know how we work in NoticiasVE.
Tags: Artificial intelligence