This topic both engages current and relevant trends with regards to AI and expands the much-discussed issue of the “invisibility of the translator” within Translation Studies. On the one hand, scholarly explorations of MT in Translation Studies have largely focused on its effects on productivity, costs, and the quality of translations it produces (Poirier & Roy 2023, Jia et al. 2019, O’Brien 2022, do Carmo 2020), or on the ethics of MT (Taivalkoski-Shilov 2018, Bowker 2020, Kenny & Winters 2020). On the other hand, the (in)visibility of the translator has mainly been explored in relation to the prevailing practices of marketing, reading and evaluating translations, to the ambiguous legal status of translation, and to the self-effacement of translators within the text of the translation itself (Venuti 1995, Porter 2006, Coldiron 2012), without sufficiently engaging with recent technological developments. This session therefore proposes to bridge the gap between the critical study of MT and scholarly debates on the (in)visibility of the (human) translator in Translation Studies. It will undoubtedly contribute to fostering ideas and dialogues in the emerging field of the study of AI within Translation Studies.
The session includes four presentations of 12 to 15 minutes each, which allows for about 30 minutes of unstructured discussion at the end. Shorter presentations allow for a more dynamic panel, with more of a roundtable feel with the discussion at the end.
In “The Translator's Game: A Ludic Approach to Machine Translation”, Anna Schewelew (PhD student, University of California, Santa Barbara) will draw on Vilem Flusser's apparatus theory in order to rethink the relationship between human translators and contemporary Neural Machine Translation systems, suggesting that agonistic play with MT can provide ways of defying the invisibility of the translator and reclaiming the power to define what translation is in the age of AI.
Michael Shea (Assistant Professor, University of Louisiana at Layafette) will argue that recent developments in Large Language Models paradoxically recenter the translator’s agency in the process of translation in his paper “Bad Translation”. Focusing on various forms of “bad translation” (from Juan Gelman’s Sydney West poems to the deeply problematic Araki Yasusada hoax), he will suggest that within a paradigm of play, AI and other machine language tools make translation a more accessible practice for future students and literary inventors.
As for Audrey Canalès (Assistant Professor, Université de Sherbrooke), she will focus on language-centric AI technologies and their impact on the film industry, to determine what works, what does not, and if human creativity is definitively obsolete. Her paper “Recent advances in language-centric AI and their impact on the audiovisual industry” will explore how AI has been increasingly used in subtitling processes, with well-known challenges such as accessibility, linguistic variation, and productions in languages other than English.
Finally, in her paper “Translation, Imitation, Authorship: After Lorca, After ChatGPT”, Rhiannon Clarke (PhD student, Johns Hopkins University) will explore the question, underlying any translation, of the authorial aura through poet Jack Spicer’s free and fake translations of poems by Federico García Lorca. Considering that the mark of a successful translation is usually that the translator’s own traces are effaced, she will argue that the translator’s defense against machine erasure is to assert her own authorial voice.
This session incorporates a variety of perspectives, as it includes two PhD students as well as two junior faculty members. Composed of three women and one man, the session also represents four different universities, and includes a Canadian scholar.