To investigate the cues that influence word learning in children with ASD, and whether this is influenced by medium of presentation (e.g., iPad or book), we worked with 16 minimally- verbal children with ASD (M receptive language: 3.9 years; M nonverbal IQ: 57.5)—the target audience for producers of communication apps on the iPad. All children were recipients of picture-based interventions such as PECS and were frequently exposed to iPads in educational settings. Participants were taught the names of unfamiliar objects presented in photographs across four within-subjects conditions: (1) via an iPad, repeatedly presenting a single representation of the target object, (2) via a picture book, repeatedly presenting a single exemplar, (3) via an iPad, presenting multiple differently colored representations of the target object, and (4) via a picture book, presenting multiple differently colored representations. Children were then tested on their ability to extend the newly-learned names to three- dimensional (3-D) referents matching on shape and color, and to generalize names to novel category members matching on shape but not color. Crucially, our results revealed that medium of presentation—iPad or book—did not impact on children’s extension of names from pictures to real objects. Rather, children with ASD only extended labels to depicted objects at above- chance rates when presented with multiple differently-colored pictures of the target referent, and tended to map narrow associative word-picture relations when presented with a single exemplar.