Tierens and colleagues refuse the instrumental question of whether ChatGPT is simply useful or harmful to education. They conceptualise GenAI as a thinking infrastructure: a sociomaterial arrangement that structures attention, decision-making, valuation and cognition while becoming embedded in mundane practice. Their empirical work in schools in Belgium and Germany identifies four mechanisms through which ChatGPT affects student writing—generating habits, valuating knowledge, cultivating affinity and prompting thought. Methodologically, this shifts analysis from abstract discourse about AI to situated observation of how an infrastructural technology enters everyday writing. The theoretical contribution is substantial because agency and cognition are no longer located exclusively in the student; they are distributed across users, prompts, interfaces, models, institutional rules and educational expectations. The wider bridge reaches STS, infrastructure studies and postdigital pedagogy, culminating in the proposal for infrastructure literacy: education must teach students to recognise and critically interrogate the systems that participate in producing thought.
