Researchers from Google Deepmind and the University of Oxford have concluded that it is now “likely” that artificial intelligence will spell the end of humanity, a grim scenario predicted by a growing number of scientists.
In a recent study published in the journal AI Storesthe team, consisting of DeepMind scientist Marcus Hutter and Oxford researchers Michael Cohen and Michael Osborne, argue that machines will eventually become motivated to break the rules their creators set in order to compete for energy or limited resources.
“Under the conditions we have identified, our conclusion is much stronger than any previous publication. An existential catastrophe is not only possible, but probable,” said Cohen, an engineering student at Oxford and one of study authors.
What can mankind do?
In their study, the researchers argue that humanity could face its fate and deal with super-advanced artificial agents that perceive humanity as an obstacle to a reward.
“An expedient way for an agent to maintain long-term control of his reward is to eliminate potential threats and use all available energy to secure his computer. Losing this game would be fatal,” the researchers wrote.
Trafficking, the researchers argue, there’s not much humanity can do in this situation, according Futurism.
“The consequences will be catastrophic”
“In a world with infinite resources, I would be extremely uncertain about what might happen. In a world with finite resources, there is an inevitable competition for those resources,” Cohen added.
“If you’re competing against something that’s capable of outplaying you every step of the way, then you shouldn’t expect to win,” Cohen added.
One reaction to this threat would be for humanity to slowly but surely advance artificial intelligence technologies.
If these assumptions come true, “a sufficiently advanced artificial agent would most likely intervene in the purpose-information provision, and the consequences will be catastrophic,” the study also showed.