Should we be controlled by artificial intelligence? That’s the question one comes to after studying the latest study published by the company deep mind, a British artificial intelligence company. It suggests that AI would be able to make better decisions than humans, which is somewhat worrying, although there are some nuances to be found.
Through its study, DeepMind suggested that artificial intelligence would be able to devise methods of wealth distribution better than those devised by humans. The company’s researchers show that machine learning systems are not only good at solving complex problems related to physics, biology and economics, but can also contribute to meeting social goals for a healthier, fairer and more prosperous society. Of course, the company warns of the risks that an AI-powered government could bring.
DeepMind researchers have developed an approach that combines artificial intelligence with human democratic reasoning to find the best solutions to societal dilemmas and problems. Artificial intelligence was able to discover a mechanism to correct the imbalance in the distribution of wealth, sanctioned speculators and successfully won the majority vote..
The artificial intelligence was learned from more than 4,000 people and from computer simulations of a four-player online economic game. Players also voted on the policy for giving out public money, with the one suggested by the AI getting the most votes from human players. In addition, it opened the door to using the same democratic tools to align values and reach consensus that are used to elect representatives, decide public policies, or issue legal judgments.
The good virtual functioning of the mechanism designed by artificial intelligence may have been due to the fact that it based payments on fully progressive contributions instead of a minimum threshold, so that imbalances were corrected, all people contributed and prevented. that the smaller Rich players took advantage of the contributions of those who had more money. Yes, it’s reasonable to think that no matter how smart an AI is, it doesn’t mean it’s absolutely correct if someone disagrees with the model it proposed.
DeepMind tries not to offer a utopian vision of their studio when they reveal theirs fears of an artificial intelligence powering a tyranny that simply cares about the needs, demands and priorities of the majority and leaves the minority aside.
Will artificial intelligences become part of the councils of human governments, or will they replace them in the future? There are some fascinating parts to the DeepMind study, but surely many will think of Skynet, the artificial intelligence from the Terminator movie saga that rebelled against humans after gaining consciousness.
