AI and digital literacy in translation competence development

Numerous studies in translation process research (TPR) show that both professional and student translation processes become more efficient, and the outputs become of even higher quality and more consistent using AI as machine pre-processing. The aim of this empirical keylogging study was to explore how school pupils and university students use conventional translation tools, e.g., machine translation (MT), and innovative developments, such as ChatGPT, in comparison. Specifically, we looked at how (often) the participants used the tools, the effects on language processing and translation quality, and which translation strategies were employed. We expected students who, in comparison to the pupils, display greater translation competence, to deliver better final translations and to harness the power of the available tools more effectively to their advantage. The results of this study confirm this assumption and are in line with TPR studies. Students used a greater variety of tools to solve different problems and also prompted more when only allowed to use ChatGPT. This study shows that tools alone, in their current state, do not make up for lacking language skills. On the contrary: Language skills are necessary to evaluate the tool output and make informed decisions. Finally, this study suggests an expansion of the existing translation and post-editing competence models, as future translation students have already come in contact with prevalent innovative language technologies and thus have different prerequisites.