'DIY AI'? Practising kit assembly, locating critical inquiry
This paper presents a reflexive ethnography of ‘DIY AI’ underway. Part 1 examines a promotional video of Google’s ‘AIY Vision Kit’, its ‘do-it-yourself intelligent camera’, running on a Raspberry Pi computer and fitting into an 4.7´7.5´7.6 cm cardboard box. Part 2 of the paper, in turn, reports on our initial effort at kit assembly with the help of the user manual. In particular, I shall home in on our ‘turn it on’ attempt, as a first ‘step condition’ to operate the assembled kit ‘intelligently’—that is, for ‘experiment[ing] with image recognition using neural networks’ (Google 2018). The reflexive ethnography pursues two aims. First, it shall make explicit (some of) the ‘vulgar enabling practices’ (Button and Sharrock 1995) of the probed ‘intelligent camera’. Second, the ethnography will revisit the interplay between ‘technical work and critical inquiry’ (Lynch 1982) by locating how, when, and why the former invited the latter in situ. Recent reflection on critical inquiry in and across STS (e.g., Mirowski 2020), algorithm studies (e.g., Mackenzie 2017), and social and cultural studies more broadly (e.g., Tsilipakos 2018), will be practically indexed and recast accordingly. So will critical inquiry in matters of ‘DIY (AI)’ more specifically.