In the interconnected world of today’s globalization, the demand for translation services has significantly risen due to cross-cultural communication, international business transactions, and online content distribution. As a result, machine translation (MT) systems have become increasingly popular for efficiently translation needs. However, the lingering issue of translation errors remains a considerable concern. This study uses news to compare the translation errors in machine translation (MT) texts generated by Instagram Translation Features and Google Translate. The study employs a descriptive comparative qualitative approach to examine translation errors in captions obtained from @Radioandikas account. The captions were translated using Google Translate (GT) and Instagram Translation Features (IGT). Data collection involved documentation of the translations from GT and IGT. The analysis involved identification, classification, calculation, and explanation to compare the translation errors between IGT and GT. The error typology presented by the American Translator Association (ATA) was utilized for the framework analysis. The findings revealed that IGT exhibited ten types of errors, while GT displayed six types of errors. These results indicate a higher variety of errors in IGT compared to GT. The study emphasized the significance of evaluating translation errors in machine translation tools to understand their limitations and make informed decisions when utilizing them for translation purposes. This study found 25 errors made by IGT. The errors were divided into 5 errors (20%) of literalness, 10 errors (40%) of usage, 1 error (4%) of omission, 1 error (4%) of syntax, 1 error (4%) of spelling, 1 error (4%) of cohesion, 1 error (4%) of faithfulness, 1 error (4%) of unfinished translation, 2 errors (8%) of verb-form, 3 errors (12%) of misunderstanding. In contrast there were 13 errors made by GT. The errors were divided into 2 errors (15.4%) of literalness, 5 errors (38.5%) of usage, 1 error (7.8%) of omission, 1 error (7.8%) of style, 1 error (7.8%) of verbform and 3 errors (23.1%) of misunderstanding. Based on the result, IGT exhibits higher translation errors compared to GT and the most dominant errors made by both IGT and GT were usage errors.

Laila , Nur .   (2023). Google Translate vs Instagram Translation features: a comaprison of translation errors. IAIN Kediri