Consciousness and the concept of internal mental models are foundational topics in neuroscience and psychology. Yet, we do not understand them well enough to engineer artificial lifeforms that are conscious. In this talk, I will be discussing a line of work on developing “world models” for artificial agents. Such world models construct an abstract representation of the agent’s world that helps it navigate in its environment. We will be discussing the use of world models and attention as a form of bottleneck for an artificial agent, connecting this line of work with ideas and techniques from computational evolution and artificial life. The goal of the talk is to encourage the development of artificial life that incorporates a form of internal mental model, which will be a stepping stone for creating conscious machines.