The present study investigates the role and effectiveness of AI feedback assistants in L2 writing instruction in tertiary education. It will show in which ways AI feedback can complement feedback by human instructors. The focus will be on student uptake and specific aspects of language instruction that readily lend themselves to AI assistance. A collaborative approach to providing feedback involving both human instructors and AI assistants in the area of planning and developing lines of argumentation will be introduced. Based on the results of a previous study (Eibinger, Fromm, Reitbauer, 2024) in which AI feedback has proven beneficial in the context of process writing, it will be argued that AI feedback assistants can provide clearer and more elaborate explanations of complex aspects of writing such as faulty lines of argumentation or ineffective supporting of arguments than teachers. AI chatbots provide their feedback often in the form of lists of bullet points. Students frequently improve their lines of argumentation upon receiving this type of AI-generated feedback. It was found that AI assistants’ feedback also contributes to teacher well-being by helping instructors to better manage their time and cognitive resources. However, a cautious approach to delegating feedback on lines of argumentation to AI exclusively is suggested since overreliance on AI assistants can weaken the development of essential critical thinking skills in students. Critical appraisal of AI generated text will be one of the most vital skills needed in the future. Students need to be able to evaluate the suitability of AI-generated feedback. This process of verification demands active and critical engagement with the chatbot, making a certain degree of critical awareness essential for students.

Keywords: AI feedback assistants, language education, EFL writing, higher education.