Home The Engine Room Engine Room Seeing without knowing: limitations of the transparency ideal and its application to algorithmic accountability