A Comparative Analysis of Google Translate and ChatGPT for the English Translation of Aranyakanda

The translation of Sanskrit texts into English, particularly religious and philosophical ones like the Ramayana, presents significant challenges due to the cultural and linguistic complexities involved. Modern AI tools have demonstrated significant progress in translating various languages by preserving their semantic depth and philosophical nuances. Leveraging semantic analysis techniques to evaluate the quality of translations will provide deeper insights into how well these models can preserve the essence of the original text. The framework presented here publishes the translation of the selected chapters of Aranyakanda using Google Translate and ChatGPT and then compares the same with the manual translation published by GitaPress using semantic analysis via a BERT-based language model. The study indicates that Google Translate is not yet reliable for complex literary translations, often misrepresenting meaning or context whereas ChatGPT can complement Gita Press translations by providing modern readability without compromising the essence of the text. This approach can contribute to improving the accuracy of translations and ensuring that the rich cultural and philosophical meanings embedded in Sanskrit texts are maintained.