We currently have ascii-art and graphical-output tags. There doesn't seem to be any tag which is unambiguously for challenges which involve output in an image file format, so they sometimes get put in graphical-output and sometimes (e.g.) don't get any tag. Moreover, some challenges which ask for ASCII-art output (e.g. 1, 2) are tagged as graphical-output but not ascii-art.
Inconsistent usage is obviously a small problem, which can be fixed by agreeing a standard and then editing tags as appropriate. But I think it's also a problem that tag completion from "image" doesn't send you to "graphical-output".
I am tempted to create an image tag and retag everything in graphical-output as either ascii-art or image, or both where the question gives the implementer a choice between the two (e.g.). However, the name doesn't strike me as adequate, because there have been questions which take images as input but don't produce them as output.