This article develops a Derridean account of generative AI as an ‘algorithmic pharmakon’ that unsettles comforting oppositions between human and machine writing in education. Drawing on Derrida’s concepts of pharmakon, supplement, iterability, archive, and the ‘university without condition,’ it argues that large language models do not invade a previously pure sphere of authorship but intensify the technical, citational, and institutional conditions that have always organised educational writing. The paper first re-reads current AI panics as a repetition of long-standing anxieties about writing, then analyses how generative systems function as supplements that both support and displace pedagogical labour. It situates AI within power-laden digital archives, examines its implications for authorship, assessment, and academic integrity, and explores how chatbots reshape attention, desire, and the political stakes of curriculum. Rather than endorsing simple bans or celebrations, the article proposes a Derridean ethics and pedagogy oriented to process, disclosure, justice, and critical work on and with AI as an unstable support for study. It concludes with practical recommendations for educators navigating this algorithmic milieu.
Bilal Hamamra (2026) studied this question.