Detailed Analysis
College educators across the United States are reporting a measurable and accelerating decline in core academic skills among their students, with the pattern appearing severe enough to prompt structural changes in how courses are designed and assessed. Professors describe students who are unable to sustain attention through a full chapter of reading, who struggle to synthesize arguments across multiple sources, and who cannot extract meaning from complex texts through sustained engagement. High school teachers echo these observations in the domain of writing, noting that the collapse in quality extends beyond the relatively straightforward problem of AI-generated submissions — even students who are not using AI tools have demonstrably lost the habit of working through a difficult draft. The distinction educators are drawing is between willful avoidance and genuine incapacity, captured in the stark phrasing that students "can't do it anymore" rather than "won't."
The response from faculty has become increasingly pragmatic and, in some ways, diagnostic of how seriously the problem is being taken. A growing number of instructors are abandoning take-home assignments entirely, deeming them functionally worthless as measures of student capability in an environment where AI assistance is ubiquitous and difficult to detect or police. In their place, in-class writing, oral examinations, and live demonstrations of reasoning are being substituted — methods that require students to perform cognitive work in real time and without technological mediation. This structural shift represents a significant departure from decades of pedagogical convention and signals that at least some portion of the academic community views the problem not as a temporary disruption but as a fundamental alteration in what students are able to do.
The article frames the phenomenon in generational terms, identifying the current cohort of students as the first to have grown up with capable AI tools available before they had the opportunity to build the underlying cognitive infrastructure those tools now substitute for. This is a meaningful distinction. Prior technological disruptions in education — calculators, the internet, search engines — arrived after students had already developed baseline skills in arithmetic, research, and information evaluation. The concern being raised here is that AI assistance arrived early enough in students' formative years to preempt the development of those foundational competencies rather than merely augment them. Skills like tolerating ambiguity in a difficult text, constructing an argument across multiple drafts, and synthesizing disparate sources require deliberate practice over time; if that practice is consistently bypassed, the underlying capacity may not develop at all.
The narrator's personal response — sitting children down with pencils and paper as a deliberate corrective — reflects a broader parental and cultural anxiety that is likely to intensify as the academic consequences become more visible. This kind of household-level intervention underscores that the concern has migrated beyond professional educational circles into the lived experience of families navigating what AI access means for child development. The explicit rejection of nostalgia as a motivation, and the framing instead as a direct response to documented evidence, positions the reaction not as reactionary technophobia but as a rational adjustment to observable outcomes. Whether such individual correctives can meaningfully offset systemic exposure remains an open and deeply uncertain question.
The broader trend this article reflects is one of the most consequential and underexamined questions in contemporary AI development: the degree to which cognitive offloading to AI systems, when adopted before underlying skills are formed, produces not enhanced humans but dependent ones. Researchers and policymakers have devoted considerable attention to questions of misinformation, bias, and labor displacement, but the subtler question of what sustained AI assistance does to human cognitive development — particularly in children and adolescents — has received comparatively little rigorous investigation. The testimony coming from educators, accumulated informally but consistently, may represent one of the earliest large-scale empirical signals about that question, arriving not from a laboratory but from classrooms where the effects are already visible and pronounced.
Read original article →