Acceso usuarios
The translation landscape is changing significantly with the advent of genAI-powered chatbots like ChatGPT. Machine translation (MT) systems have been at the forefront of translation technology since the 1950s, but the new genAI models represent a transformative shift. This study explores the advantages and limitations of translating English to Portuguese and Spanish using Machine Translation (MT) systems and ChatGPT. MT systems have improved significantly over the last years, and large language models (LLMs), democratized recently by ChatGPT, have also recently proved to be effective at translating text. Nevertheless, these systems still face challenges in accurately capturing different languages’ cultural specificities, nuances, and complexities. This paper analyzes translations performed by various chatbots and MT systems, including Systran, Google Translation, DeepL, Chat AI GPT (free), Bing Chatbot (powered by ChatGPT), and Google's Bard. The data analysis is presented from a cognitive and contrastive perspective, following the author's recent methodology. The aim is to evaluate their overall performance and effectiveness. The results reveal that translating from English to Spanish and Portuguese presents significant challenges for all systems and tools, including grammatical and semantical issues. Additionally, notable differences have been observed between the translations performed by the MT-Systems and those from genAI chatbots. Finally, this paper will discuss the advantages and limitations of MT-translation systems and their implications for the role of human translators.
