The Use Of Artificial Intelligence In Improving Machine Translation Post-Editing; Insights From Translation Editors

Abstract

The intricate nature of the human language has been a huge challenge to machine translation, and the human editors are constantly required to evaluate and access the quality and accuracy of machine translated contents. The post-editing processes have been manually conducted prior to the advancement of artificial intelligence. This study explores the roles of artificial intelligence in machine translation post-editing, focusing on gaining insights from professional translation editors on how AI models have facilitated their functions. The study is a survey quantitative analysis, and data was collected through the use of questionnaire that was distributed digitally. The survey community include 193 professional translation post editors drawn across different locations. Analysis was carried using relevant statistical measures, and the findings were presented in different statistical tables. The main AI tools employed in machine translation post editing include SDL Trados Studio, MemoQ, MateCat, KantanMT, Lilt, and MemSource. The data indicates that Lilt is the most frequently used AI tools by the study samples. The findings indicate that the integration of AI tools with the efforts of human translation editors yields excellent result in machine translation post editing. The study further indicated that majority of the participants rejected the position that machine translation post-editing tools can effectively produce high quality machine translation post-editing services without the input of human editors. There is also an indication that the human translation editors are concerned about the security of their job as advancements in AI continues to threaten their profession. However, almost, 94% of the respondents accepted the fact that the best practice is the combination of the efforts of human editors and AI machine translation post-editing tools.