Artificial intelligence will be difficult to control in the future, research
Oxford: Researchers have warned that artificial intelligence technology is likely to take a form that will be difficult to control and have devastating consequences for humanity.
Scientists fear that out-of-control artificial intelligence systems will become more dangerous once they learn that they can break the rules set by their creators.
Artificial intelligence will reportedly reach a point where it will be forced to compete for limited energy or resources, researchers from Google and the University of Oxford wrote in the AI magazine journal last month. These advanced agents can kill humans as obstacles to their goals.
The paper adds that a better way for an agent to keep his goals under control for the long term is to eliminate potential threats and use all available energy to protect his computer.
The paper explained that defeat in this game would be fatal.
“In these circumstances, we have shown that our conclusion is stronger than any previous study,” said Oxford University researcher Michael Cohen.
Michael Cohen conducted the research with his Oxford colleague Michael Osborne and Marcus Hutter, a senior scientist at Google’s DeepMind AI Lab.