The eye at hand: when visually impaired people distribute ‘seeing’ with sensing AI

Abstract

What does it mean to see and look? Can seeing and looking be done without using the eyes? This article contributes to studies in visual communication through empirical visual research into human vision, visual impairment and computer vision technologies, using video-ethnographic methods. These topics essentially enable a respecification of the concept of vision, the role of visual and material culture, and the exploration of visual means of communication in social and cultural worlds. In this article, the authors advance the understanding of visuality and vision by showing empirically how ‘seeing’ and ‘looking’ are not uniquely human abilities, but rather informational phenomena that can be achieved through distribution with a technological, nonhuman sensing AI. This study draws on video-recorded data in which visually impaired persons (VIPs) use a smartphone with a computer vision-based app while grocery shopping in a supermarket. Based on video ethnography, ethnomethodology and conversation analysis (EMCA), they show the orderly, practical organization of four specific practices: (1) fleeting glancing; (2) searching; (3) identifying; and (4) locating. In the examples, these ordinary human practices for achieving visual information are done while using the technology as a handheld ‘eye’. This research contributes to studies in visual impairment, visuospatial organization and the use of AI consumer products in a context of cultural practices for accomplishing the act of looking at and picking up grocery products. The article contributes new knowledge on visuality by expanding the concept of distributed perception and by suggesting a praxeological respecification of achieving visuospatial perception as an action in the world.