Science after Wittgenstein: Reading Encounters with the Affective Grammar of AI-Mediated Science and Educational Orientations

Abstract Artificial intelligence (AI) has attracted growing attention in science and education, yet the affective register through which scientific concepts, judgements, and norms are handled when AI becomes part of ordinary scientific work has received less examination. Rather than treating AI as an external force acting upon science, this paper attends to how scientists’ hesitation, vigilance, and quiet recalibration make the grammar of scientific practice publicly available, particularly where established ways of going on no longer run automatically. Drawing on Wittgenstein’s later philosophy, we set out a methodological orientation that approaches scientific literature as a series of reading encounters in which argument, style, hesitation, and affect render the criteria of scientific practice perceptible. What follows are three sets of reading encounters of scientists’ commentaries in Nature and Science . The first stays with AI in science as an enterprise, where AI is folded into routine descriptions of method. The second turns to medical science, where judgement, trust, and responsibility become especially pressing under AI-mediated conditions. The third takes up science, art, and creativity, not as evidence of disciplinary rupture but as occasions in which the boundaries of scientific practice are tested and repaired in use. Across these encounters, affective signals are not incidental to analysis but conditions of noticing that anything has shifted at all. The final section considers what the reading encounters could mean for science education: shifting attention away from evaluating AI as beneficial or harmful to cultivating affectively attuned judgement and the capacity to go on responsibly within evolving scientific practices.