For years humanity has struggled to effectively allocate resources and money to those who need it, leading to wars, waste, and endless political debacles.

However, all that could change in the future. According to new research by the wizzes at DeepMind, artificial intelligence has been demonstrated to have potentially better decisionmaking skills than humans when it comes to solving society's problems.

DeepMind's study saw a group of people play a game where they each receive varying levels of money and have to decide whether to keep the cash or use different methods to share the wealth. They then had to vote on which method of splitting up the profits they preferred.

READ NEXT: FBI and MI5 spooks give unprecedented joint warning on China's spying and 'tech theft'

A 'Democratic AI' was then trained on data from these games and taught to play the game itself. It settled on a 'liberal egalitarian' method which forced a minimum contribution from poorer players but shared money from the richer players.

This ensured that poorer players couldn't 'freeload' off of richer players, but also stopped richer players from stealing all the wealth for themselves.

Human players consistently voted for this AI-designed method as the fairest way of divvying up resources and addressing inequality.

The researchers argued this proved AI could be used to design 'fair and prosperous societies', and said: "The ultimate goal of AI research is to build technologies that benefit humans⁠—from assisting us with quotidian tasks to addressing grand existential challenges facing society."

• 

  YouTube lists creepy horror clips as child-friendly – despite creator's protests

However, they also said that applying AI on a large scale in this way, such as in a country's government, would be very challenging to implement, and it's not clear whether it would work.

Players involved in the study weren't told that an AI designed the mechanisms they were voting on, and so there may be less popular support for a similar model in the real world.

The researchers also identified issues with 'tyranny of the majority' which could see flawed policies introduced by an AI that have negative consequences for individuals or minority groups.

READ MORE:

• Weasel broke the Hadron Collider and sent Earth to alternate reality, bizarre theory claims
• Disabled man forced to sell late wife's gadgets and Star Wars collection to pay bills
• Mark Zuckerberg gives staff harsh warning as Meta faces 'worst downturn in history'
• AI can predict crime a week in advance with more than 90% accuracy, scientists claim
• Las Vegas hotels to offer 'first ever' VR porn delivery robots as part of room service

• Artificial Intelligence
• Boris Johnson

Source: Read Full Article