The Effect of Artificial Intelligence Generated Translation versus Human Translation on Reading Comprehension of the Speakers of Less Commonly Taught Languages

English as a Foreign Language (EFL) University students usually write in their native tongue and translate it into English using Artificial Intelligence programs. The study evaluated the quality of translations generated by AI in one hand and investigated the impact of Artificial Intelligence Generated Translation (AIGT) on EFL students in another. A human translator and an AI tool were used to translate two sample texts from English into Persian. The texts were given to ۳۰ EFL teachers to examine the quality of AI translations. In addition, ۱۵۲ students randomly divided into control or experimental groups were exposed to translated texts. Results from an independent t-test showed that there was a negligible difference between the two groups. The qualitative analysis of the interview data that involved ۳۰ participants revealed that language teachers perceived omission, addition, syntax and punctuation errors in AIGT as adequately acceptable, despite their prevalence. However, a majority of the teachers were dissatisfied with AIGT’s accuracy in rendering idiomatic expressions. Based on the results, EFL educators should acknowledge the prevalence and usefulness of AI among students, and aim to incorporate it effectively in their teaching instead of prohibiting its use.

Keywords:

Text analysis , Reading Comprehension , AI , less commonly taught languages , Persian