From Perception to Action: Can UI Interventions Foster Sustainable LLM Chatbot

Problem
The paper addresses the gap in literature regarding the sustainability of large language model (LLM) chatbots, particularly focusing on user-interface (UI) interventions to foster energy awareness. While existing strategies primarily target model and infrastructure efficiency, the potential of UI design to influence user behavior remains underexplored. This work is a preprint and has not undergone peer review.

Method
The authors conducted a baseline survey with 77 participants to gauge awareness and receptiveness to sustainability-oriented UI interventions. They developed a web-based chatbot prototype featuring a three-mode switch (Energy-efficient, Balanced, Performance), per-response energy feedback, pre-send energy estimates, a usage metrics dashboard, and energy analogies. The prototype was evaluated in a five-day field study with 11 participants, focusing on how these UI features influenced user behavior regarding energy consumption.

Results
In the baseline survey, 94.8% of participants reported some awareness of AI energy use, but 88.3% misestimated actual energy consumption. Despite high concern for environmental impact, only 39.0% were willing to accept a performance trade-off for lower energy use. In the field study, the Energy-efficient mode was selected for 55.8% of logged prompts, and 90.9% of participants reported actively choosing Eco-mode when high accuracy was not critical. Notably, participants did not reduce prompt length, indicating that mode switching was the primary behavioral mechanism driving energy-conscious interactions.

Limitations
The authors acknowledge that the small sample size in the field study (11 participants) may limit the generalizability of the findings. Additionally, the reliance on self-reported data could introduce bias, as participants may overstate their energy-conscious behaviors. The study also does not explore long-term effects of UI interventions on user behavior or the potential impact of different demographic factors on energy awareness.

Why it matters
This research highlights the importance of UI design in promoting sustainable practices in LLM chatbot usage, suggesting that user behavior can be influenced through thoughtful interface interventions. The findings indicate that sustainability-oriented UI features can enhance user awareness and encourage energy-responsible interactions without sacrificing usability. This work lays the groundwork for future research on energy-aware conversational AI design, as it provides a prototype and replication package for further exploration. The implications of this study are significant for the development of more sustainable AI systems, as published in arXiv.