LLM-Generated Explanations Nudge Sustainable Consumer Choices
A new study from arXiv (2607.25726) investigates how large language model (LLM)-generated recommendation explanations can nudge consumers toward sustainable choices. Researchers applied nudge theory to craft sustainability-aware explanations, validated through human evaluation and LLM-as-a-judge audits. Two randomized experiments (N=529) were conducted in low-involvement (instant coffee) and high-involvement (hotel bookings) domains. Participants chose among preference-matched recommendations accompanied by these explanations. Results indicate that behavioral framing of sustainability information in explanations significantly influences user choices and perceptions. The study demonstrates that generative AI can effectively embed nudges into recommendation systems, promoting more sustainable consumption without sacrificing user satisfaction.
Key facts
- Study published on arXiv with ID 2607.25726
- Investigates LLM-generated recommendation explanations as behavioral nudges for sustainability
- Uses nudge theory to design sustainability-aware explanations
- Validates explanations through human evaluation and LLM-as-a-judge audits
- Conducts two randomized studies with total N=529 participants
- Low-involvement domain: instant coffee
- High-involvement domain: hotel bookings
- Participants choose among preference-matched recommendations with explanations
Entities
Institutions
- arXiv