Every day there are more experts who warn about the danger of artificial intelligence. Or to be more precise, the evolution of this. While people like Bill Gates or Sam Altman are delighted with their possibilities, many others are the least concerned. And they insistently repeat that time runs, not precisely in our favor.
Among the latter is Charbel-Raphaël Ségerie, Executive Director of the French Center for the Safety of AI. Of course it is not someone as media as Elon Musk, for example, but it has been sufficient to dedicate themselves to AI to pay attention to his words. A few words that coincide with the pessimism shown by people like Geoffrey Hinton.
The evolution of AI, a budding danger
As explained in recent statements, Charbel-Raphaël Ségerie is clear that artificial intelligence is not on the right track. «It is a real madness,» he says, making his position very clear. And why? Basically because, according to him, It is too uncertain technology so as not to have more thorough control over it. Something very dangerous.
The problem, as almost always, is money. The desire to get it, says Ségerie. From their point of view, the big companies, starting with OpenAi, responsible for Chatgpt, only advance in the complex land of the AI with an end: earn the maximum possible money. Something that, sooner or later, can end up being a problem for all humanity, so in general.
«I am not contrary to innovation,» says Ségerie. «I’m an engineer, I adore her. But I think that the priority for humanity is to ensure that generative AI is not our latest technology«, Judgment. And that is something that is not happening at all at the moment. If everything continues, warns, even the extinction of humanity could be a possible and even probable scenario.
Charbel-Raphaël Ségerie fears the same as many other experts in AI: that the time comes when chatbots are able to manage themselves. That is, control over them is lost. A fantastic scenario? Not at all. Taking into account how fast everything is developing with AI, he believes he can pass perfectly, and without waiting too much.
The risk of losing control over technology is real
In fact, the expert considers that it is a threat that is already happening. «Some advanced AI systems have already exhibited deceptive and harmful behaviorand, despite this, these models are being given more autonomy to make actions and decisions in the world, «it reason. It is what he understands as» self -replication. »
Or what is the same, that the AI grows and expands as a virus without human intervention. As if that were not enough, Ségerie gives the matter date. He predicts that even at the end of this year could occur, although Most likely, around 2027 becomes a reality. By then, yes, maybe it would be too late to react.
Know How we work in NoticiasVE.
Tags: Artificial intelligence