By building new instances of life, AL offers the possibility to observe independent (at least genetically, if not conceptually) examples of life without leaving the planet. Furthermore, these new instance of life may include non-carbon based life.
To talk of creating new life forms may sound preposterous or frightening. We should clarify what we mean. Unfortunately, there is no generally agreed upon definition of life, thus we can not easily talk about what it means to create life.
Generally, efforts to define life are based on testing for a list of properties. The problem arises from a lack of agreement on what should be included on the list. Properties common to many lists include the abilities to: replicate, evolve, metabolize, respond to stimuli, repair damage. Most examples of ``Artificial Life'' will fail any such test, unless the list of properties is very short.
Yet this life-test list approach is not quite satisfactory. As a provocative example of why, consider a machine that could engage as an equal intellectual participant in a discussion of these issues. Even if this machine did not replicate, or evolve, or show most of the properties that occur on most life-test lists, it would be hard to deny that it is in some sense alive.
Such considerations lead to an alternative way of approaching the problem. This approach involves making a long list of properties which are known to occur only in living systems. Now rather than asking if an example of AL exhibits all items on the list, we ask if it represents a genuine instance of any item on this list. If so, then we can way that we have captured a genuine instance of some property of life in our synthetic system.
Most work in AL has this kind of flavor. The author of the work is generally interested in some aspect of life, such as evolution, intelligence, language, social behavior, development, etc. The author then creates a system, which if successful, will exhibit life-properties in the author's area of interest. Yet the system may exhibit none or few of life's other properties. So they become in some way dis-embodied examples of life. It may not be appropriate to ask if they are alive, a question which we can not answer anyway. It might be best to ask if they exhibit genuine properties of life, a question that can more easily be answered.
The field of AL operates in three main domains: hardware, software and wetware synthesis. This essay will focus on evolution, the life-process of primary interest to this author. While there has been some very exciting work with synthetic evolution in the chemical medium [13,2,7,12], and some new developments that may eventually permit the evolution of hardware [5,33], this essay will focus only on software evolution.