What Kind of Translation Literacy Will Be Automation-Resistant?

Abstract

Automated language processing is advancing rapidly, notably through neural machine translation and automatic text production using various large-language-model (LLM) generative pre-trained transformer (GPT) technologies. These are destined to have an impact on any language-intensive profession, especially translation. The nature of that impact will in turn limit and orient the knowledge and skills that translators can market effectively, both in work with the technologies and in language tasks where the technologies flounder in one way or another. The empirically based O*Net database can help identify the knowledges and skills of translators that are likely to be exposed to automation, and thereby point to those that might be expected to remain automation-resistant. That analysis in turn enables some recommendations to be advanced for the future of translator training, notably with respect to the use of spoken communication skills, the gaining and maintenance of trust, and the ability to draw on automation at key points in translation workflows in which professionals will still have a place.