Can artificial intelligence really be a danger to humanity? According to many experts, the answer is affirmative. Even its main validers, people like Sam Altman, the founder of Openai and creator of Chatgpt or Elon Musk himself, have recognized their risks. Now a new security expert has spoken about it.
It is an authorized voice, since It has experience in related security issues, precisely, with AI. In his opinion, there are some keys that will be fundamental. They will depend on them that, indeed, this technology can end up getting out of control. And without the need to resort to apocalyptic scenarios of the most famous science fiction films.
The concerns of an AI expert
The name of Chris Sestito may not be as well known as Sam Altman’s or other world experts in artificial intelligence, but the truth is that, as far as artificial intelligence is concerned, he has enough to say. More than anything, because It is the co -founder and CEO of HiddenLayera company specialized in cybersecurity which is focused on preventing attacks related to AI.
In a recent interview granted to the Axios portal, it has recognized that there is a chance that AI is a threat. A priori, it could be thought that being a part interested in the subject (he dedicates himself precisely to that) his voice has a relative validity. However, if your arguments are attended, these are sufficient precise to be very reasonable.
For the expert, the key is in the criteria that are usually used when facing the security issues that concern AI. Sestito regrets that They usually tend to confuse what he calls «algorithmic biases» and «security threats» that is certainly real. Or in other words, that the dangers of this technology are not being taken enough.
The most striking thing about his vision of the matter happens because he does not talk about a rebellion of machines as such (he does not discard it completely), but of much more practical and accurate cases. That is, even today they can be given. For example, Technical problems directly related to AI algorithms can help anyone manufacture a weapon.
The authentic dangers that artificial intelligence brings
In this way, he once insists on an issue that is key to him: vulnerability. Not only in terms of hackeos, that the specialist thinks that it seems the only thing that really worries, but in other important elements that in one way or another would have to pay much more attention.
For example, he declares that other challenges cannot be ignored that artificial intelligence has brought with it, and that now, due to this technology, They allow to extract sensitive information from discriminatory algorithmic models and decisions. If you attend what it says, AI can be very indiscreet, and that is something that should be taken.
Know How we work in NoticiasVE.
Tags: Artificial intelligence