Large Language Model Agents for Improving Engagement with Behavior Change Interventions: Application to Digital Mindfulness
Harsh Kumar, Suhyeon Yoo, Angela Zavaleta Bernuy, Jiakai Shi, Huayin Luo, Joseph Williams, Anastasia Kuzminykh, Ashton Anderson, Rachel Kornfield
arXiv Preprint Archive July 3, 2024 preprint
Study at a glance
AI-extracted from the abstract| Characteristics | Randomized experiment |
|---|---|
| Sample size | 556 |
| Population | Crowdworkers in a single-session study and participants in a three-week study |
| Duration | Single session; three weeks |
| Topics | Meditation |
| Keywords | Cs.hc Cs.ai Cs.cy |
| Key findings | Both information-providing and self-reflection-facilitating LLM agents increased intentions to practice mindfulness, but only the information-providing agent with a friendly persona significantly improved engagement with the exercises. The authors suggest that specific LLM agents may help bridge the social support gap in digital health interventions. |
Abstract
Although engagement in self-directed wellness exercises typically declines over time, integrating social support such as coaching can sustain it. However, traditional forms of support are often inaccessible due to the high costs and complex coordination. Large Language Models (LLMs) show promise in providing human-like dialogues that could emulate social support. Yet, in-depth, in situ investigations of LLMs to support behavior change remain underexplored. We conducted two randomized experiments to assess the impact of LLM agents on user engagement with mindfulness exercises. First, a single-session study, involved 502 crowdworkers; second, a three-week study, included 54 participants. We explored two types of LLM agents: one providing information and another facilitating self-reflection. Both agents enhanced users' intentions to practice mindfulness. However, only the information-providing LLM, featuring a friendly persona, significantly improved engagement with the exercises. Our findings suggest that specific LLM agents may bridge the social support gap in digital health interventions.
Comparable studies
Other randomized controlled trials on meditation, most cited first.
| Study | Year | Design | Participants |
|---|---|---|---|
| Alterations in Brain and Immune Function Produced by Mindfulness Meditation Healthy employees in a work environment | 2003 | Randomized controlled trial | n = 41 |
| Prevention of relapse/recurrence in major depression by mindfulness-based cognitive therapy. Recovered recurrently depressed patients | 2000 | Randomized controlled trial | n = 145 |
| Open hearts build lives: Positive emotions, induced through loving-kindness meditation, build consequential personal resources. Working adults | 2008 | Randomized controlled trial | n = 139 |
| Short-term meditation training improves attention and self-regulation Undergraduate Chinese students | 2007 | Randomized controlled trial | n = 40 |
| Mindfulness-Based Cognitive Therapy for Depression: Replication and Exploration of Differential Relapse Prevention Effects. Recovered recurrently depressed patients | 2004 | Randomized controlled trial | n = 73 |