Seeking a balance between the polarized positions currently coexisting in academic and professional contexts, this study explores the pedagogical potential of Generative Artificial Intelligence (GenAI) in translating popular science texts. Acknowledging the enduring relevance of the paradigmatic opposition between “apocalyptic” and “integrated” attitudes (Eco 1964), we propose a measured, critically informed approach grounded in university translation teaching practice. Our hypothesis posits that AI can render translation’s quintessential challenges—such as figurative language and cultural references—more visible and intelligible, while also fostering reflective, responsible student attitudes. Through a German – Spanish case study and utilizing the conversation forum as a means of data collection, we highlight the contrastive capacity and stimulating value of AI, provided it is deployed under instructor guidance within quality-oriented tasks. The results are articulated in twelve emerging theses that advocate for a cautious, supervised integration of AI and aim to stimulate pedagogical debate within translation didactics.