It is typically assumed that translation learners should just become familiar with various text types relevant to the course throughout the semester, the fact which leads to the marginalization of other important aspects of translation. This study aims to propose a more purposeful teaching approach to translation training. For the purpose of this study, the course in translating scientific and technical texts was selected. At the beginning of the course, a pretest consisting of technical texts with varying levels of difficulty was administered to the learners. Translations were assessed using the Waddington’s (2001) evaluation model, and the most frequently encountered translation issues were then identified. Subsequently, a similar test was conducted, and students were asked to translate using an AI-powered translation tool. Translations were again assessed based on the Waddington’s model, and translation problems were identified. Given that the use of AI-based translation tools is essential for students to remain competitive in the current translation market, only those issues that emerged in translations in both tests were chosen as the basis for instruction. Therefore, the instructor sought to select texts throughout the course that included specific features likely to cause the same translation problems identified in the pretests, allowing students to practice repeatedly so that their weaknesses could be effectively addressed. This research highlights the importance of administering a pretest for needs assessment in translation teaching. It also indicates that, given the emergence of artificial intelligence and its application in translation training, students do not encounter some of the issues that they would face in the past when AI-powered tools were not as widespread.
