The Death of the Student Essay–and the Future of Cognition (www.forkingpaths.co)

A college professor argues that generative AI—exemplified by ChatGPT—has effectively “killed” the student essay, substituting sparkling, error‑free prose for the messy cognitive work of thinking and writing. Grading 95 papers, he found two patterns: genuinely thoughtful student work, and a new genre he dubs “glittering sludge” — fluent but hollow text often containing fabricated citations or misattributed sources. Because large language models (LLMs) can polish grammar, invent plausible but false references (hallucinations), and will soon be able to autonomously research and draft entire assignments, detecting AI use is becoming a losing battle and rubs faculty down to “citation police.” The essay warns this is more than a grading problem: it threatens the cultivation of “how to think,” undermines the link between assessment and actual learning, and risks offloading core human cognitive skills to machines. Technically, the danger stems from LLMs’ fluency and emerging agentic capabilities, which make authored output indistinguishable from student-produced reasoning. Practical responses include shifting assessments (e.g., proctored in‑person exams), redesigning pedagogy to emphasize process over final prose, and intentionally integrating AI where it augments rather than supplants critical thinking. At stake, the author contends, is language as humanity’s superpower—the practice of thinking made visible through writing—and our ability to preserve it in an AI era.