The use of generative chat in education has become widespread over the last four years, raising many questions about its use and the effects of AI on learning. The aim of the current systematic review is to analyze the main effects of feedback through the use of generative chat on the production of academic texts by university students. This research is defined as a systematic review of the literature according to the guidelines of the PRISMA statement. The search was conducted in three international important databases (Scopus, Eric, and WoS), from which 12 articles were selected. The results highlighted that there are positive effects on university students’ writing when generative chat is used as a means of providing feedback. Among the main results, it was observed that feedback via chat helps to improve aspects mainly associated with the structure and organization of texts, allows for the proper use of grammatical conventions, and improves the fluency and cohesion of sentences, as well as the precision of ideas and vocabulary. In addition, other benefits were observed in the review, such as improved self-efficacy, self-regulation, proactivity, motivation, and reflection on writing, which promotes critical thinking about the text but also about AI, reducing anxiety and stress.