The central driving force of intelligent agents will be based on economic reward and punishment. The reward is the increase in the market value of digital objects held and managed by intelligent agents and the punishment is the loss of market value of those assets. 

The market itself will provide the input necessary to determine which new combination of information should survive and which should not. The economic constraints will mean that the process of generating new information will be as efficient as possible.

https://key.studio/#/?id=agi-model