METM26 presentation

Digital misaddress and power in automated social media translation

Nancy Tahir, Lisbon, Portugal

Automated translation on social media platforms introduces new forms of digital misaddress, where messages are redirected and reframed in ways that shift power dynamics between sender and receiver and reshape the participants’ identities in the discourse. This presentation, based on original research undertaken as part of my doctoral studies, examines this phenomenon and its impact on power. Power refers to the authority that enables the speaker to direct the narrative and influence the course of events. Misuse of such power or loss of its control is represented in the hidden algorithmic intervention that turns the speaker from a position of credibility and strength into one of suspicion or exposure to legal consequences or political exploitation.

In this presentation I will analyse a sample of case studies from different social media platforms such as Facebook, WeChat, and Instagram using a qualitative comparative discourse analysis; the data is based on posts that have been translated from languages such as Arabic, Hebrew, and Thai into English, and vice versa. The analysis identifies several types of potential errors, including technical limitations represented by context collapse in translation and systemic bias when different cultural norms are imposed.

This presentation has significant implications for translators, editors, transcreators, and content creators as professionals; it transforms their role from being text producers into guardians of digital context and linguistic risk consultants. The presentation gives language professionals a glimpse into how they can assist their clients with preventive authorship and strategic post-editing, which would be very helpful for politicians, activists, public figures, and governmental representatives who are affected by such errors.

This research aims to contribute to a new understanding of the relationship between language, technology, and power in the digital environment, highlighting the value of human-in-the-loop interventions that lies in spotting ‘algorithmic gaps’ which can cause digital misrepresentation.

About the presenter

Nancy Tahir is a certified legal (Arabic<>English) translator and court interpreter. She holds a Master’s in Translation from Saint Joseph University, Dubai, and is currently a PhD candidate in Translation Studies at Universidade Católica Portuguesa. Her interests include legal and literary translation, AI, neural MT, MTPE, ethics and human–AI collaboration.