Aplicativos móveis de interpretação automática: expectativas e percepções dos usuários brasileiros

• Ano de defesa: 2022
• Autor(a) principal: Freitas, Flávio de Sousa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Estudos Linguísticos
• Palavras-chave em Português: interpretação automática; expectativas e percepções dos usuários; tradução automática; speech-to-speech translation; users' expectations and perceptions; machine translation; CNPQ::LINGUISTICA, LETRAS E ARTES::LINGUISTICA::LINGUISTICA APLICADA
• Link de acesso: https://repositorio.ufu.br/handle/123456789/35038; http://doi.org/10.14393/ufu.di.2022.5018

Despite the numerous systems developed over the nearly four-decade history of speech-to-speech translation, little is known about the expectations and perceptions that users have about this service. The present study seeks to identify the perceptions of Brazilian users of speech-to-speech services offered by mobile applications from the App Store, the Apple's mobile application store. And, from this identification, this study briefly analyzes the user’s expectations. The results show that several comments shared by the users convey contradictory expectations and perceptions and, no matter how much the user separates them into periods, the fact that these expectations and perceptions compose the same comment can lead the observer to have difficulties with their interpretation. Users' expectations and perceptions point to the need for applications to improve the services offered by them, especially regarding the translation of texts in images offered by Google Translate and the offline translation functionality offered by Voice Translator. In addition, the data show that user expectations and perceptions deserve attention from both speech-to-speech app developers and human interpreters, as the parameters that must be considered when judging the quality of automatic translations and interpretations are not always within the reach of that user who has never consumed that kind of service.