A cellular automaton is a grid of cells, each in one of a few states, that all update together in discrete steps according to a simple rule based only on a cell’s own state and the states of its neighbors. There is no central controller and no global plan; every cell follows the same local rule. The concept was introduced by John von Neumann and Stanislaw Ulam in the late 1940s and early 1950s, when von Neumann was looking for a clean mathematical setting in which to prove that a machine could reproduce itself.
What makes cellular automata important is the gap between how simple the rules are and how rich the resulting behavior can be. From a few local rules you can get stable structures, moving patterns, periodic cycles, and even configurations that perform computation. Von Neumann showed that a cellular automaton could contain a universal constructor able to build any machine described on a tape, including a copy of itself. Later work demonstrated that some cellular automata are capable of universal computation, meaning they can in principle compute anything a general computer can.
For the prehistory and ongoing study of artificial intelligence, cellular automata are a clean illustration of emergence: complex, lifelike, even intelligent-seeming behavior arising from many identical simple units following local rules, with no designer specifying the overall outcome. That principle, simple rules plus many interacting parts giving rise to complex behavior, recurs in neural networks, artificial life, and self-organizing systems. The primary source used here is Burks’s report reconstructing von Neumann’s automata.
For a general reader, cellular automata are the cleanest demonstration that you do not need a master plan to get complex behavior; the right simple rules, repeated everywhere, can be enough.