Animating AI: language variation and “enoughness” in voice user interfaces

Jones, R. ORCID: https://orcid.org/0000-0002-9426-727X (2026) Animating AI: language variation and “enoughness” in voice user interfaces. Annual Review of Applied Linguistics, 46. ISSN 1471-6356 (In Press)

Abstract/Summary

This paper develops an interdisciplinary perspective for understanding how AI chatbots project recognisable social identities and how these performances shape human interactions with them. Bringing together concepts from critical sociolinguistics, anthropology, philosophy, and HCI and UX research, it argues that large language models function as technologies that animate cultural identities by activating links between linguistic behaviours and social identities. The perspective brings together Silvio’s notion of animation, the collection of social practices through which “humanness” is projected onto non-human entities, and Blommaert’s notion of enoughness, the idea that the authenticity of linguistic performances is not so much a matter of the accuracy of a performance as it is a matter of how audiences collectively evaluate bundles of semiotic cues as sufficient, insufficient, or excessive in ratifying a performance as socially recognisable. The analysis draws on a corpus of metapragmatic artefacts related to ChatGPT’s advanced voice mode and Sesame.ai’s hyper-realistic voice interface posted on social media sites like TikTok and Reddit. Analysis of these artefacts reveals how designers, users, and AI systems co-produce boundaries of authenticity through the deployment and uptake of enregistered linguistic and discursive features such as accent and stance. In doing so, they continually recalibrate what counts as culturally competent performances, shaping emergent norms of identity and sociality around AI. The paper highlights how humanness and culturality are distributed across technical systems, corporate discourse, and human interlocutors, with important implications for understanding how generative AI reproduces cultural stereotypes by drawing on the linguistic labour of users.

Item Type	Article
URI	https://centaur.reading.ac.uk/id/eprint/130486
Refereed	Yes
Divisions	Arts, Humanities and Social Science > School of Literature and Languages > English Language and Applied Linguistics
Publisher	Cambridge University Press for the American Association for Applied Linguistics
Download/View statistics	View download statistics for this item