Artificial Life programs break down into two basic categories: Cellular Automata (CA) and Genetic Algorithms (GA). This new form of "life" is electronic, and exists only within the confines of a computer. Although many, many orders of magnitude simpler in design than anything found in nature, ALife forms can be used to model biological organisms and ecosystems, or solve incredibly complicated optimisation problems.
Cellular Automata repeat a series of rules over and over again, applying them to produce amazingly sophisticated patterns. One famous CA is John Horton Conway's "Life", which produces shapes which "walk" across the screen, and explode in pulsating patterns. One slightly fantastical line of thought is that the entire universe is one complex CA, obeying a set of rules with precise mathematical precision.
Genetic Algorithms are programmes which have a "genetic code" built into them. Many of these programmes are released and they behave as their (initially random) genes dictate. After a while, the computer assesses which of the programmes have been best at meeting a "goal" (such as "solve this equation") and "breeds" the successful ones together. After a number of generations, natural selection will have developed programs which do their task very efficiently. This can be used to develop instructions for a robot to follow when avoiding obstacles, or in finding an effective antibody against bacteria.
For more information, ask Google.