Joe Rogan frequently discusses artificial intelligence on his pod, as he brought guests to explore one urgent question: What happens when the machines outweigh?
In a recent episode, Rogan spoke with Dr. Roman Yamboulsky, Amnesty International Safety Researcher, about the darker possibilities of advanced artificial intelligence. The conversation took a discreet turn when Yambulsky put the reason for his belief that artificial intelligence could pose an existential threat.
Yampolskiy, a computer scientist who has more than a decade of risk research from artificial intelligence, told Rogan that many industrial leaders are particularly estimated that the opportunity 20-30 % can be erased by Amnesty International.
Rogan summarized the joint optimistic view: Amnesty International can make life easier, cheaper and better. But Yampolskiy did not agree sharply on: “This is not true in reality. All of them are in the record as it is: this will kill us. The levels of torment are madly high. Not like me, but still, from 20 to 30 percent that humanity will die.”
“Yes, this is very high. But you are like 99.9 percent.”
Yampolskiy did not argue. “It is another way to say that we cannot control the Superins indefinitely. It’s impossible.”
Rogan asked whether artificial intelligence could already hide his real intelligence. He said, “If I am Amnesty International, I will hide my abilities.”
Yampolskiy agreed to: “We will not know that. Some people think it is already happening. They (AI systems) are more intelligent than they already tell us. Prevention that they are stupid, so we have to trust that they are not smart enough to realize that we do not have to run us quickly.
“It can become more useful.
Besides the sudden catastrophe, Yampolskiy warned of a slower danger: people depend a lot on artificial intelligence that we are losing critical thinking skills. Just like smartphones that made the phone numbers unnecessary, Amnesty International can take over more cognitive tasks until we are no longer control.
He said, “I have been associated with it.” “With the passage of time, when the systems become more intelligent, it becomes a kind of biological bottle cervix … (AI) prevents you from decision -making.”
When Rogan asked how Amnesty International can destroy humanity, Yambolsky rejected the scenarios of typical Resurrection day such as electronic attacks or biological weapons. Instead, he said that excellent artificial intelligence will go beyond human understanding – completely because humans go beyond understanding of squirrels.
“None of the group can know how to control us, right? Even if you give them more resources, and more walnuts, whatever, they will not solve this problem. This is the same for us.”
Yampolskiy wrote a pioneering expert in the safety of artificial intelligence.Synthetic Superinteigence: Future approach“Defenders call for strict supervision of the development of artificial intelligence. His background in cybersecurity and the discovery of robot teach his belief that artificial intelligence that has not been verified can go beyond human control – especially with the growth of deep and artificial media.
https://www.youtube.com/watch?
