In 1986 the creationist chemist Charles Thaxton used the term “specified complexity” when claiming that messages transmitted by DNA in the cell were specified by intelligence, and must have originated with an intelligent agent. The Intelligent Design concept of “specified complexity” was developed in the 1990s by mathematician, philosopher, and theologian William Dembski who said that when something exhibits specified complexity, one can infer that it was produced by an intelligent cause (i.e., that it was designed) rather than being the result of natural processes.)
Dembski defines “complex specified information” (CSI) as anything with a less than 1 in 10150 chance of occurring by (natural) chance. Critics say that this renders the argument a tautology: complex specified information cannot occur naturally because Dembski has defined it thus, so the real question becomes whether or not CSI actually exists in nature.
The conceptual soundness of Dembski’s specified complexity/CSI argument has been widely discredited by the scientific and mathematical communities. John Wilkins and Wesley Elsberry characterize Dembski’s “explanatory filter” as eliminative, because it eliminates explanations sequentially. They argue that this procedure is flawed as a model for scientific inference because the asymmetric way it treats the different possible explanations renders it prone to making false conclusions.
Richard Dawkins, another critic of Intelligent Design, argues in “The God Delusion” that allowing for an Intelligent Designer to account for unlikely complexity only postpones the problem, as such a designer would need to be at least as complex. Other scientists have argued that evolution through selection is better able to explain the observed complexity.