Read for us ... and show us the pictures (Version 2.0)

Based on an earlier installation, Read for us … and show us the pictures, which debuted at ISEA 2015, The Readers Project presents the work of a software entity that generates digital video montage, with visual content sourced through live image search.
The Montage Reader analyses its text and first establishes a overall visual grammar based on closed-class words that underlie linguistic structure.
The reader then searches for images corresponding to phrases – ‘longest common phrases’ whenever possible – finally composing a sequence of still and animated images and video, that corresponds with the written language of the text both structurally and also semantically – at least in so far as contemporary image search proposes a correspondence that is meaningful for the human user-readers of network services and their aggregation of crowd-sourced indexing. The chief text read by the Montage Reader is ‘Some Thing We Are,’ a short story by Daniel C. Howe.

(Source: Artists' statement)