Przymusinski’s Autoepistemic Logic of Knowledge and Belief (AELKB) is a unifying framework for various non-monotonic formalisms. In this paper we present a semantic characterization of AELKB in terms of Dynamic Kripke Structures (DKS). A DKS is composed of two components - a static one (a Kripke structure) and a dynamic one (a set of transformations). Transformations between possible worlds correspond to hypotheses generation and to revisions. Therefore they enable to define a semantics of insertions to and revisions of AELKB-theories. A computation of the transformations (between possible worlds) is based on (an enhanced) model-checking. The transformations may be used as a method of computing static autoepistemic expansions.