While most people are wary about AI’s making decisions, a recent study from a Google AI showed that the technology can be beneficial. In the study, the AI was tasked with solving wealth inequality, and it might be better at distributing wealth than actual humans.
The current social class system has always been a problem but we haven’t seen much progress in making that better. If this Google AI can really solve this problem, maybe humans should listen to AI more.
How the Google AI solved distribution woes
Researchers at Deepmind (via Singularity Hub) were able to have a Google AI come up with new distribution methods by using a game where each player has to share their resources for mutual benefit. Due to how the game is designed, other players will eventually do better, leading to some being rich and others being poor.
In the end, Google's AI settled on a method where players would get nothing unless they contributed roughly half their private wealth. Apparently, the AI was able to fare better than other humans who played the game because it based payouts on contributions that were relative rather than absolute.
This supposedly helps with the wealth imbalance, as forcing a minimum contribution prevents less wealthy players from simply free-riding on the contributions of wealthier ones. It's an interesting direction but one that might not have the full backing of actual humans.
AI still isn’t perfect
Despite some of the impressive decisions made by the AI, researchers at Deepmind have pointed out how difficult it is for these life forms to understand justice and fairness. The fact that it took a game to help with their understanding shows that there’s still a long way to go before AI can make decisions like this regarding fair distribution.
The researchers also pointed out how their method for the AI to come up with this solution has its own faults. Using a four-player game to judge how society can distribute its wealth fairly is one thing but researchers also pointed out how players of the game weren’t told when they were playing with the Google AI, which can make humans trust them less.
Overall, we hope a compromise can be made so that humans and AI can help each other out.