Researchers from the European Max Planck Society have concluded that superintelligent AI cannot be controlled. This is stated in the report of scientists, which is published in the Journal of research of artificial intelligence.
Scientists analyzed the ability of humanity to control superintelligent artificial intelligence, according to Interesting Engineering, and prevent its attempts to end humanity. Now the conclusion is disappointing — the current level of technology development makes humanity powerless against super-intelligent AI.
To prevent the destructive actions of AI, scientists consider it necessary to create a so-called “content algorithm” that will simulate the dangerous behavior of superintelligent AI.
Study co-author Manuel Sebrian, head of the digital mobilization group at the Max Planck Center for People and Machines, noted that for now, “thead-intellectual machine that controls the world sounds like science fiction.” Instead, “there are already machines that perform certain important tasks on their own” and programmers do not know “how they learned it.”
Sebrian stated that at some point intelligent machines can become “uncontrollable and dangerous to humanity.”
Recall, the developer of artificial intelligence Jeffrey Hinton, who in 2012 created a neural network, resigned as vice president of Google. The scientist is concerned that new technologies of companies can surpass human intelligence. With its release, the “godfather” of AI warns humanity about the dangers of artificial intelligence.
