Although some published studies have challenged the importance and efficacy of written corrective feedback (WCF), ESL/EFL tutors continue to value its practice and provide WCF to their students in spite of their time and workload constraints. The development of various automated-feedback applications over the last few years contributed to our understanding of what machine-operated feedback is capable of in comparison to teachers. The release of OpenAI's ChatGPT in 2022 has taken the world by surprise. This AI-operated tool has attracted the attention of students, teachers and researchers in the field of L2 teaching and learning. There are now claims that it has the capabilities for providing WCF on written output. The overarching aim of this study is to compare the quantity and quality of analytical corrective feedback provided and moderated by trained classroom teachers with ChatGPT-generated feedback on the written output of a group of students (English Level CEFR B1) attending an EAP course at an international branch campus of a UK university. Both teachers and ChatGPT used the same analytical marking criteria to mark students' essays and provide feedback on language use and use of academic vocabulary and style. Analysis of results provided valuable insights into the adequacy of the feedback strategy, guidance and polarity, offering a foundation and guidelines for future endeavours to optimize the integration of AI-generated feedback within L2 contexts. Implications and recommendations regarding the potentials/future of using ChatGPT as a feedback tool in L2 are discussed.