If you run a computer program that models the weather, most people would agree that it is just a simulation. There is no real weather inside your computer. But what happens when you start your word processor? Is it really a word processor or just a simulation of a word processor? Most people would say it was real, not a simulation. So, how about AI or artificial life: real or a simulation? In the case of artificial life, how do we know whether or not something is simulated or real? Is there a point at which a model of life becomes so life-like that it ceases to be a model and becomes alive? A new paper (PDF format) by Jean-Philippe Rennard proposes some new ways to think about this and suggests that a paradigm shift may be needed before we can really answer the questions involved.