Artificial Intelligence is without a doubt one of the greatest technological advances of the last century. Such is the case that its use has become natural in the daily lives of millions of people and even in the jobs of thousands of people.
It is increasingly integrated into all human points of view.. However, there are many questions that the experts, creators, disseminators and technicians specialized in AI ask themselves daily in the face of
Is it possible to stop Artificial Intelligence?
In a video broadcast on social networks he explains that: “The best thing for society would be to stop Artificial Intelligencebut this cannot be stopped if we do not put in place a transparency system where we can all know that the other has stopped, because I cannot allow myself to be ahead of myself.”
“You see Elon Musk saying in a podcast that “if I could I would stop this, but it’s not up to me”, the CEO of Anthropic saying “we have to stop this for the good of society”, the CEO of Google saying “hey this has to be done well and not quickly” and then nobody does anything, it’s a capitalist race“.
To this, he adds other statements “I have many doubts that we have the remote capabilities to stop it.“I think that even if we wanted to, we don’t have the legal mechanisms in case one day we want to press the red button, but I don’t think we have them. And it’s going to take longer to agree on developing that than it will take AI to advance to a point that is not stoppable.”
The first attack carried out by an AI
According to the Spanish Data Protection Agency, the attacking agent started a search for vulnerabilities in generic files, and logged in correctly. Once he accessed the system, he began to search, autonomously, for vulnerabilities in the application, which, once achieved, allowed him to modify personal data and access invoices.
Prior to any type of conclusion, It should be noted that the information available comes from the notification submitted by the affected organization and must be subject to the corresponding analysis.
In this way, the use of a specific AI model does not imply that its provider’s model or infrastructure has been compromised nor that the tool has been designed to carry out malicious activities.
Now, what is relevant from the perspective of data protection is that a third party would have used an AI agent as an instrument to successfully chain together different phases of the attack.
This first notification does not allow us to affirm a statistical trend, although it does constitute a significant signal that attacks supported by artificial intelligence are no longer a theoretical risk and are beginning to materialize in incidents that affect real processing of personal data.
He National Cryptological Center (CCN) has also warned about this evolution. In its guide to good practices against offensive AI, the agency points out that these capabilities are already being integrated into real campaigns and raises the need to strengthen vulnerability management, identity security, supply chain supervision and control over the use of artificial intelligence agents.