La posedición de materiales informativos del ámbito sanitario en la era de la inteligencia artificial: comparación de ModernMT y ChatGPT-3.5
Post-Editing of Informational Materials in the Healthcare Sector in the Age of Artificial Intelligence: A Comparison of Modernmt Lite and Chatgpt-3.5

Summary/Abstract: Adapting appropriately to a particular client can mean the success or failure of a translation assignment. In this regard, computer-assisted translation (cat) tools have provided significant digital support which, thanks to the skills and competencies of professional translators, facilitates effective communication beyond linguistic borders. Post-editing is not new, but it remains widely used, as it allows for greater control over machine translations and their results, significantly improving the effectiveness of the workflow. The aim of this research is to evaluate the post-editing eff ort by simultaneously comparing the performance of two machine translation engines, Modernmt Lite and the language model Chatgpt-3.5, in translating informational materials for an ngo. The results demonstrate that the conversational language model based on artificial intelligence gpt-3.5, not being a machine translation tool per se, achieved better results than the Modernmt Lite model. Although the complete replacement of human labour is not yet feasible, this is only the beginning of opening a new Pandora’s box in terms of artifi cial intelligence, with possibilities and challenges that are difficult to predict.

• Details
• Contents