Artificial Life (ALife) is a growing field that brings together research from several areas. Its subject is defined loosely as the study of ``life as it could be''  as opposed to ``life as it is'' -- which is the the subject of biology.
Work in ALife includes robots that emulate animal behaviors, agents that survive in virtual worlds, artificial evolution, reinforcement learning, autonomous robotics and so on. There is a continuing debate on this field regarding what the definition, methods and goals of Artificial Life are.
We propose that one of the fundamental goals of ALife research is to be a constructive approach to the problem of emergence of complexity. Not satisfied with a global description which describes the process through abstract elements, ALife should consider the question settled only when those elements have been formalized up to the point where they can be laid down in the form of a computer program and shown to work by running it.
Whereas evolutionary biology looks at the fossil record and tries to describe the evolution of life as a result of the g-t-r dynamics, Artificial Life research should aim at writing g-t-r programs that show how in fact artificial agents increase in complexity through the process, thus proving that natural complexity can be generated by this formal process.