Recognizing carbon-saving impacts increases the likelihood of selecting eco modes in LLM interactions¶

Claim · All claims
Evidence · 1 study · design r2 · q2 · i3 large
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i3 Recognition of carbon-saving impacts was the strongest positive predictor of eco-mode selection (OR = 3.879, p < .001 when modeled alone; OR = 5.016 in the combined model). → Hyeonwook Kim 2026

Evidence¶

Hyeonwook Kim 2026¶

Hyeonwook Kim, Xuesi Chen, Alex Cabral, Cindy Kaiying Lin, Udit Gupta, Josiah Hester. (2026). When LLMs Slow Down: How Environmental Impacts Mediate University Students' LLM Usage. https://arxiv.org/abs/2608.23968 doi:10.48550/arxiv.2608.23968

q2 · i3 · design · r2

In the same GLMM analysis of the live-interface study, "recognition of carbon-saving impacts emerged as the strongest positive predictor of eco mode selection" with OR = 3.879, p < .001; the combined model reported OR = 5.016 for carbon recognition.

"recognition of carbon-saving impacts emerged as the strongest positive predictor of eco mode selection, increasing the likelihood of eco preference both when modeled alone (OR = 3.879,p < .001)"

Discussion¶

Related Claims¶

• Eco-mode selection in an LLM interface declines sharply as response latency increases, dropping from about 45% of interactions in eco mode 1 to below 5% in eco mode 5 — related
• Stronger eco-mindedness raises initial eco-mode preference, but this advantage diminishes as response latency increases — related
• Interacting with an eco-feedback LLM interface surfaces gaps in students' understanding of digital sustainability and supports learning about the environmental impact of everyday LLM use — related
• User satisfaction with eco modes declines sharply from mode 1 (mean 3.64) to modes 4–5 (means 2.35 and 2.06) — related
• Concrete, outcome-oriented eco-feedback visualizations with real-world analogies are rated more favorably than abstract metrics — related