What happened? A new 'Democratic AI' has been developed that can distribute wealth better than humans. DeepMind, a UK-based AI company, has created an AI that excels in value alignment - a concept that refers to how well the values and goals of an AI system are aligned with what humans actually want. One of the key obstacles to value alignment is that human society is diverse, making it unclear to whose preferences AI should align. The team trained an AI agent for the task of wealth distribution, using real and virtual interactions between people to guide it toward a desirable outcome.
Why is this important for human beings and our society? The DeepMind team-taught an AI system to distribute the money fairly using real and virtual human interactions to help it reach a good and hopefully fairer conclusion.
The so-called "Democratic AI" studied a game known as the public goods game, in which players can invest money into a fund and earn rewards based on their investment level. Traditional redistribution methods were also evaluated, as well as an innovative method created with deep learning using the Human-Centered Redistribution Mechanism (HCRM).
The development of AI has the potential to drastically improve our lives - but only if it is done responsibly. With great power comes great responsibility, and we must make sure that AI is developed in a way that benefits everyone, not just a select few.
"We emphasize that our results do not imply support for a form of 'AI government', whereby autonomous agents make policy decisions without human intervention. We see Democratic AI as a research methodology for designing potentially beneficial mechanisms, not a recipe for deploying AI in the public sphere," the researchers added.
So what do you think? Is this a step in the right direction? Or is it a cause for concern? Let us know in the comments below!
Author: Christian Kromme
First Appeared On: Disruptive Inspiration Daily
The latest disruptive trends with converging technologies that will change your life!