Tricky
Briefly immortal
If Dembski's assertion were correct, then we should see no extinctions, especially those preceded by long, dead-end strings of evolution. For example, trilobites evolved greatly and expanded widely, but they all died, leaving no successors. I do not see how those eons of evolution followed by extinction could be called part of a directed process. It also ignores the well-established knowledge that most mutations are either benign or harmful. If it were directed, then most or all mutations would be beneficial.Dembski asserts that such systems cannot evolve in small successive steps that are not part of a goal directed process, but that isn't part of the definition. Actually, his assertion is somewhat more complicated than that, but that's an ok approximation.
Of course, this all gets swept away with the comment, "there is a 'director' but we don't know his plan." This is a cop-out pure and simple. How can you tell the difference between a plan that no-one can understand versus no plan?