This paper aims to analyze the ability of Gemini Advanced and ChatGPT 4o to translate an animated film for children from Brazilian Portuguese into Italian, taking into account the target audience (children) and a language as close as possible to spontaneous Italian speech. The translations were also compared with those of students, mostly Italian speakers, of Portuguese as a foreign language at B1/B2 CEFR levels. The results show that although the two systems were able to produce good translations, they still produced incorrect, inaccurate or context-inappropriate translations requiring human intervention. Furthermore, the preliminary results of this research may serve as a guide for the teaching of Portuguese as a foreign language and pragmatic-discursive aspects in audiovisual translation from a contrastive perspective.
