The Ghost of the Author: Mechanisation, Metricisation and Narrative Contraction in the Age of AI

Abstract

This article examines the crisis of scientific writing in the age of artificial intelligence (AI) not as a sudden, isolated technological rupture, but as the inevitable consequence of a long-term structural transformation. Arguing that AI has not created this crisis but rather made pre-existing contradictions visible, the study grounds its argument on four primary axes: the historical rhetorical construction of scientific discourse, the transformation of the IMRAD (Introduction, Method, Results, Discussion) format into an unquestionable normative model, the profit-driven political economy of academic publishing monopolies, and the increasingly ambiguous status of authorship in the context of large language models. Drawing on Michel Foucault’s conceptualisation of knowledge-power, Thomas Kuhn’s approach to normal science, and Peter Fleming’s critique of academic capitalism, this analysis reveals how the current system rewards only measurable and rapidly circulable outputs while marginalising theoretical risk-taking and critical depth of interpretation. The article highlights that the discrepancy between detected and declared AI use in academic production points to a severe transparency problem. As texts become standardised, measurability increases; AI, in turn, reproduces these forms perfectly, thereby radically narrowing epistemic and narrative diversity. Proposing three potential future scenarios – full automation, the resurgence of creative theoretical writing, and the coexistence of a dual literature structure – the study concludes that a permanent solution lies not in superficial technical fixes, but in profound structural interventions. These include the radical restructuring of the publishing economy, the elimination of academic precarity, and the revalorisation of non-standard, critical forms of writing. The critique developed here is directed primarily at the empirical natural and social sciences, where IMRAD-based writing and metric pressures are most acute, though the structural forces identified – academic capitalism, publishing monopolies, and AI optimisation – operate system-wide.