(ORDO NEWS) — Many science fiction films have featured an advanced AI that wants to “rule the world.” For example, you can see this in the well-known Terminator franchise. Most are accustomed to considering it just fantasy, which will never be, but experts have conducted another study and proved that we are all wrong.
Study author Manuel Alfonseca from the University of Madrid explained that humans, even in theory, cannot control superintelligent AI.
The situation becomes even worse, because it will be almost impossible to detect it, especially when it is on the verge of its creation. But at the same time, it is too early to worry, because people are very far from the creation of such artificial intelligence, and this can happen in several decades, or even hundreds of years.
Alfonse decided to conduct such a study after he saw the news that artificial intelligence has been beating people in some games for a long time.
After the research, he believes that it will not be possible to control AI because of the fundamental limitations that are present directly in the computational process itself. He added that an algorithm that will try to ensure that AI does not harm people must first create a model of its behavior in order to understand exactly what the consequences of its actions may be. In addition, the algorithm must stop the superintelligent machine if everything goes wrong and the AI begins to harm people.
Contact us: [email protected]