Identifying Cultural and Semantic Translation Errors in Pashto–English Proverbs Translation: A Comparative Study of ChatGPT, Google Translate and Gemini Translations

Abstract

Machine Translation (MT) has advanced rapidly with the emergence of neural and AI- powered systems, 
yet translating culturally embedded figurative language particularly proverbs continue to pose significant 
challenges, especially in low-resource languages such as Pashto. This study examines how accurately these 
three major AI translation tools such as ChatGPT, Google Translate and Gemini interpret and translate 
20 culturally rich Pashto proverbs into English. Therefore, by applying a qualitative research design 
supported by basic quantitative analysis, the current study evaluates the semantic, figurative, and cultural 
accuracy of the AI-generated translations. Moreover, the study uses Vinay and Darbelnet’s (1958) Model 
of Translation Strategies as the analytical framework for identifying the strategies employed and the types 
of errors made by each tool. The analysis of the study reveals substantial variation in performance by 
highlighting that ChatGPT achieved the highest accuracy of 75%, effectively applying equivalence and 
modulation strategies to preserve figurative meaning, Gemini performed moderately with the frequency of 
65% with occasional semantic shifts whereas, Google Translate showed the lowest accuracy of 15%, 
frequently producing literal, fragmented, or culturally inappropriate translations. The findings of the study 
highlighted the limitations of AI systems in handling proverbs from low- resource languages, demonstrating 
that figurative and culturally bound expressions require oblique translation strategies that current MT tools 
often fail to apply. The present study contributes to MT research by providing empirical evidence from a 
low resource language such as Pashto and underscores the need for culturally informed, context-sensitive 
AI translation models.