The Plagiarist in the Machine? Generative AI and the Will to Fail

This paper argues that the real challenge posed by Large Language Models (LLMs) in Higher Education lies not in their potential for plagiarism, but in their creation of a new form of writing that is indistinguishable in the traditional essay. Only through a phenomenological analysis of the writing process can the genuine difference become perceptible. We contrast the lived experience of traditional composition with that of chatbot‐assisted generation, revealing how the latter alters the writer's relation to anxiety, authorship and failure. Both the concerns surrounding LLMs and their uncritical embrace, we suggest, share a Cartesian illusion: that the mind might be freed from the body's drudgery, and that thinking could be detached from the act of writing itself along with the risk of its failure. In displacing effort and vulnerability, the chatbot offers a false promise of liberation that mirrors the instrumental logic of the contemporary university. Higher education, already redefined as the cultivation of marketable skills for individual advancement, thus confronts a deeper crisis: the obsolescence of the very practices through which it claims to form intellect and character. The essay, as both form and assessment, becomes the site of this crisis—caught between obsolescence and the diminishing will to fail.