I Quit Teaching Because of ChatGPT
Growing use of ChatGPT and other large language models in schools is forcing a rethink of what it means to learn, write, and be assessed. Many commenters worry that students are outsourcing difficult cognitive work to AI, undermining the development of critical thinking and writing skills, while others argue these tools can function as powerful tutors and productivity aids if education shifts toward in‑person assessment and higher‑order tasks. The exchange highlights deeper structural issues in schooling—large classes, busywork, shallow assignments—and a looming divide between those who use AI to augment understanding and those who rely on it as a crutch.
LLMs and student learning
- Many worry that students will offload thinking and writing to LLMs, skipping the “cognitive workout” needed to build real skills.
- Others see LLMs as powerful, patient tutors and editors that can lower barriers to learning, especially for those who struggle with traditional schooling.
- Several compare LLMs to calculators: useful once fundamentals are mastered, harmful if used before basic skills are learned. Others argue this is a false equivalence because LLMs can generate entire essays, not just answers.
- Some note that younger people already struggle with focus, reading long texts, and analytical thinking; LLMs may exacerbate this trend but are likely a symptom, not the root cause.
Cheating, assessment, and academic standards
- Many see current assignment and grading structures as brittle: anything that can be done asynchronously and alone is easily outsourced to AI.
- Proposed responses include more in‑class, pen‑and‑paper or supervised writing; low‑stakes frequent writing; and public/peer‑visible work to discourage generic AI output.
- There is concern that students can now obtain degrees with shallow understanding, as AI-generated work is “good enough” to pass.
- Others argue this simply exposes how much schoolwork was already busywork.
Teachers adapting with AI
- Some teachers report using LLMs to generate quiz questions, lesson plans, reports, and feedback, freeing time for direct student interaction.
- Others feel grading AI‑written essays is demoralizing and unsustainable, prompting thoughts of quitting.
Tools vs crutches: analogies to other tech
- GPS, backup cameras, calculators, and spellcheck are cited as precedents: tools that atrophy some skills but increase safety and efficiency.
- Counterpoint: these tools are narrow, predictable, and easily supervised; LLMs are broad, hard to detect, and can replace entire cognitive tasks.
Equity, class divides, and long‑term effects
- Some predict a split: those who use AI as an amplifier of existing skills vs those who use it as a crutch and never develop depth.
- There is debate over whether AI will democratize high‑quality tutoring or entrench new forms of dependence and control, especially given corporate ownership and opacity.