This paper presents a method to automatically generate audio-visual slideshows from text articles using word concreteness. Key words are selected based on their concreteness, and images are retrieved from searches using these words. User studies found that slideshows created with this method were strongly preferred over those using basic keyword searches, and were not significantly different than manually created ones. While effective, the method is limited to articles with concrete language and cannot identify non-famous people or objects. Future work includes incorporating video clips and other word properties.