“It Looks Good Enough”: Recognizing the Quality of Generative AI Output in Academic Writing Tasks in Higher Education

Generative Artificial Intelligence (Gen AI) is being used for writing tasks. According to previous research, Gen AI has become more reliable, yet it is prone to hallucinations, faulty reasoning, and poor referencing practices. This paper addresses the following: How can HEIs prepare their students to recognize or critique the level of output that Gen AI is producing? Eighty-nine graduate students in a hospitality management school in Switzerland entered the same prompt into ChatGPT. We posit that ChatGPT cannot create convincing graduate-level writing and students cannot identify higher-level responses. The originality of this study resides in the ability to gauge the Academic Writing (AW) output effectively. The premise: If students do not have sufficient foundational skills, they will be unable to judge whether the work produced by Gen AI is accurate or at the graduate level. Until they are trained, students should not use current Gen AI tools for AW tasks.