Imagine sitting down and playing a simple computer game against your friends where you all create a civilization by giving them parameters to act under in the world. However, unlike other simulation games like SimCity or Sid Meier’s Civilizations series you do not control how they develop. There is also only a finite amount of resources available for them use. Instead each round you are able to make choices on how you might reward your civilization and they have to balance that against other parameters they need to grow, survive and flourish. Whenever your civilization meets a certain metric you score points. It is a simple enough concept that can be created with deep learning.
Each player also has their own AI assigned to them. They work to make the players happy and progress is made across several intervals to figure out how best to do this with their citizens while each player designs and updates some sort of reward function. However, it can still be competitive among players as there are only so many resources in the world and each one of them creates their goals for their civilization without the others knowing. On its own, this is AI harmlessly interacting with people for enjoyment.
Similar things could be used to model population development among animals and also test how we as humans interact with each other. Both cooperative and aggressive mechanics can be easily built into this sort of a game and different players will choose each one. However, the overall trends that results from this is something that sociologists and game theorists might have great interest in. Additionally, the AI could theoretically be for managing anything as long with while the AI reward function creation is crowd sourced.
This would be an unethical usage of AI if the players are not clearly informed, in large print with its’ own screen not a short line in the terms and conditions that are linked to online. Although AI can be trained to solve a variety of problems it takes a great deal of time to train them and crowd sourcing their design might be more effective. The AI was collaboratively built by both the developer and the player so failing to make the player aware that they are giving this up by playing the game should be a violation of intellectual copyright even though the player is less aware of the problem they are trying to solve. Which is why players should also be made aware of that problem. Although, AI is examined for its role in creating ethical issues, it is equally worth examining if the process involved in creating AI are also ethical. All those contributing to the design should be made aware of what it is to be used for so they can decide if it is ethical to contribute to.