Gmail’s Smart Compose:  A Critical Composit(ion)

1. Abstract

While Google once merely monitored users’ words, today the company literally writes them. This is thanks to Smart Compose, a word-prediction algorithm that Google has launched in Gmail and Google Docs. The algorithm depends on Google’s meticulous recording and machine-reading of the personal data of an untold number of its 1.5 billion Gmail users, leveraging users’ words and writing for the development of the technology. In this context, language is particularly vulnerable to corporate intervention and manipulation. As such, this presentation carefully considers Fréderic Kaplan’s (2014) call to action: “through… the advent of algorithms as a new media, something is likely happen [sic] to language, and, although we are not yet sure what it will be, new tools must be built in order to understand this global linguistic evolution” (p. 62).

Responding to Kaplan's call, I report on experiments with Smart Compose in which I am manually transcribing more than 50,000 words from published texts and, subsequently, annotating and visualizing input to and output from the algorithm. These experiments are part of my larger doctoral project that seeks to locate the shifting "semantic coordinates" (Striphas, 2015, p. 398) of "language," "words," and "writing" in an algorithmic culture. Applying the framework of data colonialism (Couldry & Mejias, 2019), I argue that word-prediction algorithms such as Smart Compose must necessarily shift our understanding of these terms when words become data (Thornton, 2019) and writing becomes a datafied practice.

More broadly, I suggest that in place of the question posed by Siva Vaidhyanathan (2011) some years ago—“what do we gain and what do we lose by inviting Google to be the lens through which we see the world?” (p. 9)—we must, urgently and necessarily, ask this: what do we gain and what do we lose by allowing Google to offer the words through which we write the world?

References

Couldry, N. & Mejias, U.A. (2019). Data colonialism: Rethinking big data’s relation to the contemporary subject. Television & New Media, 20(4), 336-349. https://doi.org/10.1177/1527476418796632

Kaplan, F. (2014). Linguistic capitalism and algorithmic mediation. Representations, 127(1), 57-63. https://doi.org/10.1525/rep.2014.127.1.57

Striphas, T. (2015). Algorithmic culture. European Journal of Cultural Studies, 18(4-5), 395-412. https://doi.org/10.1177/1367549415577392

Thornton, P. (2019). Language in the age of algorithmic reproduction: A critique of linguistic capitalism [Unpublished doctoral dissertation]. Royal Holloway, University of London.  

Vaidhyanathan, S. (2011). The googlization of everything (and why we should worry). University of California Press.