Abstract meaning representation (AMR) is a graph language that encodes the semantics of natural language sentences into graphs. Natural language sentences are broken down into vertices that represent actions and entities. Named edges connect entities to actions, determining their role in the action. AMR strips away the syntactical details of natural language, meaning two distinct but semantically equivalent sentences are encoded into the same graph. AMR is used for natural language processing (NLP) tasks such as generating English text. Generating English text relies on processing large, manually annotated corpora of AMR. This work uses graph extension grammars (GEG), a graph generative formalism, to augment AMR corpora. The contribution is three strategies for manipulating GEGs, increasing their output of AMR within their domain. The aim is to make it easier to increase the size of existing AMR corpora.