Enhancing behavioral nudges with large language model-based iterative personalization: A field experiment on electricity and hot-water conservation
作者: Zonghan Li, Yi Liu, Chunyan Wang, Song Tong, Kaiping Peng, Feng Ji
分类: cs.CY, cs.AI, cs.HC
发布日期: 2026-04-07
💡 一句话要点
利用大语言模型提升行为干预的个性化效果
🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 行为干预 个性化指导 大语言模型 节能减排 认知负担 随机实验 用户参与 行为摩擦
📋 核心要点
- 现有的行为干预方法在反馈转化为具体行动时常面临认知负担,限制了其有效性。
- 本研究提出利用大语言模型生成个性化指导,并在干预过程中进行迭代更新,以降低认知负担。
- 实验结果显示,LLM个性化干预在电力节约方面表现最佳,较传统干预方式显著提高了节约率。
📝 摘要(中文)
本研究探讨了如何通过大语言模型(LLM)生成个性化指导,增强行为干预的有效性。研究在中国的233名大学生中进行,比较了三种不同的干预方式。结果显示,LLM个性化干预显著提高了电力和热水的节约效果,尤其在干预初期表现突出。研究表明,LLM的迭代个性化能够有效促进行为改变,且在行为摩擦的影响下,效果有所不同。未来需要更大规模的试验和对更多行为的扩展。
🔬 方法详解
问题定义:本研究旨在解决传统行为干预方法在反馈转化为具体行动时的认知负担问题。现有方法在不断变化的环境中,往往无法有效指导用户采取可行的后续步骤。
核心思路:本研究的核心思路是利用大语言模型生成个性化的指导建议,并通过迭代更新来适应用户的变化需求,从而减轻认知负担,提升干预效果。
技术框架:研究设计了一个三臂随机实验,分别为LLM个性化干预组、图像增强传统干预组和文本传统干预组。每组参与者在干预过程中接收不同形式的指导,干预效果通过电力和热水的节约量进行客观测量。
关键创新:本研究的主要创新在于将大语言模型应用于行为干预的个性化指导中,显著提高了干预的适应性和有效性。这一方法与传统的静态干预方式形成鲜明对比。
关键设计:在实验中,LLM根据参与者的反馈和行为数据生成个性化建议,采用了动态更新机制。关键参数包括个性化建议的生成频率和内容的上下文适应性。
🖼️ 关键图片
📊 实验亮点
实验结果显示,LLM个性化干预组在电力节约方面每日减少0.56 kWh的消耗,相较于传统文本干预组节约率提高了18.3个百分点。该优势在干预的前两轮中显著体现,并在后续保持稳定。
🎯 应用场景
该研究的潜在应用领域包括能源管理、公共卫生和环境保护等。通过个性化的行为干预,能够有效促进用户的节能减排行为,提升社会整体的资源利用效率。未来,研究成果可扩展至其他行为领域,推动更广泛的社会行为改变。
📄 摘要(原文)
Nudging is widely used to promote behavioral change, but its effectiveness is often limited when recipients must repeatedly translate feedback into workable next steps under changing circumstances. Large language models (LLMs) may help reduce part of this cognitive work by generating personalized guidance and updating it iteratively across intervention rounds. We developed an LLM agent for iterative personalization and tested it in a three-arm randomized experiment among 233 university residents in China, using daily electricity and shower hot-water conservation as objectively measured cases differing in friction. LLM-personalized nudges (T2) produced the largest conservation effects, while image-enhanced conventional nudges (T1) and text-based conventional nudges (C) showed similar outcomes (omnibus p = 0.009). Relative to C, T2 reduced electricity consumption by 0.56 kWh per room-day (p = 0.014), corresponding to an 18.3 percentage-point higher adjusted saving rate. This advantage emerged within the first two intervention rounds, alongside iterative updating of personalized guidance, and persisted thereafter. Hot-water outcomes followed the same direction but were smaller, less precisely estimated, and attenuated over time, consistent with stronger friction in this domain. LLM-personalized nudges emphasized prospective and context-specific guidance and were associated with higher participant engagement. This study provides field evidence that LLM-based iterative personalization can enhance behavioral nudging, with behavioral friction as a potential boundary condition. Larger trials and extension to more behaviors are warranted.