When LLMs Slow Down: How Environmental Impacts Mediate University Students' LLM Usage

Synthesis: Large language models' carbon and water footprints raise important Sustainability concerns, particularly with adoption rates exceeding 80% among university students despite limited insight into the environmental impacts of individual usage. Kim and colleagues design and deploy an eco-feedback interface that visualizes latency-carbon trade-offs during live LLM interactions, studying its use with 89 undergraduate computer science students in a computing Ethics course. They find that the likelihood of choosing the eco-feedback system is significantly shaped by sustainability awareness, providing an empirical look at how a technically sophisticated and values-oriented user population responds to sustainability-aware AI interfaces.

Key Findings

1. LLM adoption exceeds 80% among university students, yet there is limited insight into the environmental impacts of individual usage.
2. An eco-feedback interface visualizing latency-carbon trade-offs during live LLM interactions was designed and deployed.
3. The interface was studied with 89 undergraduate computer science students (ages 18-24) in a computing ethics course.
4. The likelihood of choosing the eco-feedback system is significantly shaped by sustainability awareness.
5. The study provides empirical evidence on how a technically sophisticated, values-oriented user population responds to sustainability-aware AI interfaces.

Connected Concepts

• LLM
• Higher Ed
• Student Experience
• Ethics
• Generative AI
• CS Education
• AI Education
• Sustainability

Connected Articles

• AIED Carbon Footprint Reporting — The Environmental Cost of LLMs in AIED
• Shen Sustainable AI Knowledge Base CS Education 2026 — Towards sustainable AI knowledge-base assistants in CS education
• GenAI Over Reliance Learning 2026 — From Enhancement to Over-Reliance
• Unesco AI Guidelines Chemical Education 2026 — Translating UNESCO AI Guidelines to Chemical Education
• Long AI Higher Ed Engagement Teaching Methods 2026 — Artificial intelligence in higher education: a systematic review

Citation

Kim, Chen, Cabral, Lin, Gupta, & Hester (2026). When LLMs Slow Down: How Environmental Impacts Mediate University Students' LLM Usage.