One of the basic paradigms underlying the rapidly expanding field of artificial intelligence states that information-processing capabilities must be interpreted as being essentially independent of the material mechanisms that embody the various performances. Within this perspective, even living systems seem to lose most of their peculiarities by being treated as one special category (among many) of information-processing "machines." The heuristic advantage of such technomorphous approaches lies in enhanced quantification: theoretical models from both fields, i.e., computer science and biology, are able to promote reciprocal understanding. We may clarify the reach of this information paradigm, however, by introducing and focusing on only a few fundamental systemic and evolutionary criteria. With applying these criteria, the biologist arrives at the conclusion that we must deny the existence of true cognitive capacities in artificial systems, whereas the computer connectionist is left with the uncertain hope of achieving one day the extraordinary robustness of living organisms via an intensified development of autoregulative simulations.