The Effectiveness of DeepSeek as a Translation Tool on Students’ Reading Comprehension

Abstract

This study aims to determine the effectiveness of using DeepSeek AI as a translation tool on students' reading comprehension. This study was motivated by the difficulties experienced by 10th-grade students at MA Unggulan K.H. Abd. Wahab Hasbulloh in understanding English texts. The study used a pre-experimental design with a one-group pre-test–post-test model and involved 35 students selected through purposive sampling. The instrument used was a 20-item multiple-choice test, which had been tested for validity using Aiken's V formula and reliability using Cronbach's Alpha. The results showed that the instrument had very high validity and strong reliability (α = 0.896). The normality test showed that the pre-test data was normally distributed, while the post-test data was not, so the analysis of score differences was performed using the Wilcoxon Signed Rank Test. The test results showed a significance value of < 0.001, indicating a significant effect in students' reading comprehension after using DeepSeek AI. These findings prove that DeepSeek AI can help students understand texts better through accurate and contextual translations, thereby facilitating the process of interpreting and understanding the content of the reading material. Thus, this study concludes that DeepSeek AI is effective as a translation tool on students' reading comprehension. The integration of artificial intelligence technology in English language learning can be an alternative solution to support the strengthening of students' literacy and reading strategies in the digital age.