Rethinking machine translation literacy for mediators: a conceptual review through participatory training with migrant teaching assistants

Purpose The academic paper reviews how a machine translation (MT) literacy programme worked and how particular design decisions, taken participatorily, shape the literacy that subsequently develops. Its case for review is drawn from participations of teaching assistants who translate on behalf of pupils, parents and institutions. Design/methodology/approach Participatory action research was implemented as an illustrative case across two training groups at learning centres in Mae Sot District, Tak Province on the Thailand–Myanmar border. Design memos, meeting notes, observational fieldnotes and follow-up reflections were used reflexively, each between-group revision being recorded together with the observation. Findings Text-operational skills proved largely absent from participants' accounts. What developed instead were judgements taken before and around translation: whether a text should be put into MT at all, which mode suits which situation and whom to consult. Trust became conditional upon topics. Checking the output was distributed among persons and bounded by time. Privacy was reframed as a duty owed to others. Practical implications Four principles are conceptually discussed and offered. Teaching time should focus on the concepts assistants understand the least. Activities should be simple enough for them to use in their own classrooms. Training materials should include examples relevant to their context. Checking should rely on local resources rather than a far-away expert. Originality/value As a contribution to Translation Studies, the participatory design concept with a case exemplar is useful for reflecting on how the finished programme can be adjusted according to contexts. Each adjustment should be done together with the person who proposed it. It also extends process-based accounts of MT literacy to mediators working with low-resource languages.