.
This scoping review, guided by the PRISMA-ScR protocol (Tricco et al., 2018), maps the literature on artificial intelligence (AI) applied to sign language machine translation (SLMT) published between 2019 and 2025. Drawing on 87 studies from Scopus, Web of Science, IEEE Xplore, and ACM Digital Library, the review employs a Translation Studies (TS) framework anchored in Jakobson’s (1959/2000) intersemiotic translation and Derrida’s (1967/1997) critique of phonocentrism. The findings reveal a field that predominantly treats sign language translation as a pattern recognition problem rather than a translational act involving cultural and multimodal meaning-making: 62.1% of the studies are gloss-based, 72.4% rely primarily on BLEU, and 90.8% do not engage with TS. These patterns describe a text-centric field, which the article then interprets, following Bauman (2008), as a computational instantiation of phonocentrism, termed algorithmic logocentrism. The review identifies critical gaps in gloss-free approaches, Deaf community participatory design, and engagement with TS theory. A research agenda for Intersemiotic AI Translation Studies is proposed, with implications for the Brazilian context and Libras.
Authors
- Bruno Galasso (ORCID: https://orcid.org/0000-0002-3677-7650)
Institutions
- Universidade Federal de São Paulo (BR)
Publication Details
- Journal
- Ilha do Desterro A Journal of English Language Literatures in English and Cultural Studies
- Published
- 2026-11-16
- DOI
- https://doi.org/10.5007/2175-8026.2026.e111420
- Primary Topic
- Translation Studies and Practices
- Type
- article
- Field-Weighted Citation Impact
- 0.00