Equivalence in Artificial Intelligence Technologies Generated Turkish and English Translations of “Little Red Riding Hood”: The Case of ChatGPT

Artificial intelligence (AI) technologies are becoming increasingly effective in translation studies as in many other disciplines and fields. Although the development of AI assistants has led to translations closer to the source text compared to previous years, it is still possible to encounter errors. It is thought that most of these errors arise from a lack of detailed information provided to the AI assistant regarding the type of source text and/or the target reader of the target text. For example, with fairy tale translations, the lack of data such as rules that should be followed about the genre and the demographics of the target reader increase the number of errors. The aim of this study is to examine with Werner Koller’s translation theory of equivalence the ChatGPT-generated German to Turkish and English translations of “Little Red Riding Hood”. For this purpose, ChatGPT was instructed to translate the German source text into the two target languages and the extent to which the translations met the criteria of denotative, connotative, text-specific, pragmatic, and formal equivalence was analysed. The analyses reveal that, with the Turkish translation, ChatGPT uses a grammatical tense inappropriate for the fairy tale genre; several words or phrases are mistranslated, and translation errors arise from a failure to consider cultural differences. The English translation also shows that several words are mistranslated, but other equivalence criteria are met more accurately compared to the Turkish translation. Based on these findings, it is recommended that AI assistants be given more detailed instructions regarding the genre of the text and the demographics of the target audience such as age, gender and culture to improve translations that meet equivalence criteria.