The issue of Testability has been brought up by many people. Even conservative columnist George Will brought it up in his critique of Intelligent Design.
Here is William Dembski's response :
Your deeper concern is that intelligent design is not science because it is not testable. If ID were not testable, you would have a point. But the fact is that ID is eminently testable, a fact that is easy to see.
To test ID, it is enough to show how systems that ID claims lie beyond the reach of Darwinian and other evolutionary mechanisms are in fact attainable via such mechanisms. For instance, ID proponents have offered arguments for why non-teleological evolutionary mechanisms should be unable to produce systems like the bacterial flagellum (see chapter 5 of my book No Free Lunch [Rowman & Littlefield, 2002] and Michael Behe’s essay in my co-edited collection titled Debating Design [Cambridge, 2004]). Moreover, critics of ID have tacitly assumed this burden of proof — see Ken Miller’s book Finding Darwin’s God (Harper, 1999) or Ian Musgrave’s failed attempt to provide a plausible evolutionary story for the bacterial flagellum in Why Intelligent Design Fails (Rutgers, 2004).
Intelligent design and evolutionary theory are either both testable or both untestable. Parity of reasoning requires that the testability of one entails the testability of the other. Evolutionary theory claims that certain material mechanisms are able to propel the evolutionary process, gradually transforming organisms with one set of characteristics into another (for instance, transforming bacteria without a flagellum into bacteria with one). Intelligent design, by contrast, claims that intelligence needs to supplement material mechanisms if they are to bring about organisms with certain complex features. Accordingly, testing the adequacy or inadequacy of evolutionary mechanisms constitutes a joint test of both evolutionary theory and intelligent design.
Unhappy with thus allowing ID on the playing field of science, evolutionary theorist now typically try the following gambit: Intelligent design, they say, constitutes an argument from ignorance or god-of-the-gaps, in which gaps in the evolutionary story are plugged by invoking intelligence. But if intelligent design by definition constitutes such a god-of-the-gaps, then evolutionary theory in turn becomes untestable, for in that case no failures in evolutionary explanation or positive evidence for ID could ever overturn evolutionary theory.
I cited earlier Darwin’s well-known statement, “If it could be demonstrated that any complex organ existed which could not possibly have been formed by numerous, successive, slight modifications, my theory would absolutely break down.” Immediately after this statement Darwin added, “But I can find out no such case.” Darwin so much as admits here that his theory is immune to disconfirmation. Indeed, how could any contravening evidence ever be found if the burden of proof on the evolution critic is to rule out all conceivable evolutionary pathways — pathways that are left completely unspecified.
In consequence, Darwin’s own criterion for defeating his theory is impossible to meet and effectively shields his theory from disconfirmation. Unless ID is admitted onto the scientific playing field, mechanistic theories of evolution win the day in the absence of evidence, making them a priori, untestable principles rather than inferences from scientific evidence.
Bottom line: For a claim to ascertainably true it must be possible for it to be ascertainably false. The fate of ID and evolutionary theory, whether as science or non-science, are thus inextricably bound. No surprise therefore that Darwin’s Origin of Species requires ID as a foil throughout.