Traducción automática en contextos clínicos: los cuestionarios de autodiagnóstico

This article analyses the weaknesses and risks of machine translation (MT) in clinical contexts where doctor-patient communication is relevant by focusing on the importance of the translation and adaptation of self-diagnostic questionnaires. The analysis of the MT of the ChEAT questionnaire identifies several limitations in purely linguistic terms, and in pragmatic and sociocultural terms, which result in a lack of adaptation to the target audience, namely children. These limitations affect the quality of communication in medical settings, and stem from the MT's inability to approach all the elements of the translation process from the outset, which ultimately leads to inaccurate results and perpetuated biases. The translation of scales is a multi-step process aimed at quality assurance. The need for human intervention in clinical contexts and the danger of replacing human translation with machine translation is therefore underlined.