One of the structural problems of modern societies, especially in Latin America, is the problem of the concentration of wealth, inequality and the misuse of resources, something that has sparked protests and crises in several countries in the region. And while politicians and businessmen have yet to join forces to resolve the issue, a Artificial Intelligence (AI) is working on it.
Google DeepMind researchers have released a piece of paper in which they explore the possibility that an AI can create a more egalitarian society by dividing resources more equitably.
The study reveals a series of experiments in which a neural network is responsible for dividing resources equally, and as humans prefer. To do this, the participants played the public goods game online, a type of game based on experimental economics. In this game, a series of tokens or coins are distributed equally, and in each round, the player decides whether to keep them in a private and individual fund, or contribute to a public fund.
The funds are then distributed through three different redistribution schemes based on existing economic systems. In addition to these three schemes, an additional one is added, created entirely by an AI and which they called ‘Human Centered Redistribution Mechanism’ (Human Centered Redistribution Mechanism) or HCRM. After, the participants had to choose which type of system they preferred.
In most of the times, the participants chose the system generated by the AI.
Pursuing a broadly liberal egalitarian policy, [HCRM] sought to reduce pre-existing income disparities by compensating players in proportion to their contribution.”
“In other words, rather than simply maximizing efficiency, the mechanism was progressive: It promoted the right to vote for those who started the game with an economic disadvantage, at the expense of those with greater initial wealth.”