Toward less air travel: evidence from a multi-component field experiment combining GenAI and peer-based digital nudges

This is the authors' abstract. We don't add key points for this paper.

Reducing carbon emissions from holiday air travel—an important contributor to overall emissions in Western countries—is crucial for the transition to a post-carbon society. This article presents short-term findings from a randomized controlled trial (RCT) conducted in Switzerland, examining the effect of a digital nudging intervention aimed at reducing demand for holiday air travel. In 2025, 204 self-selected volunteers were randomized to control and treatment groups; 146 completed both baseline and first follow up surveys over a three-month intervention period. Rooted in the Model of Action Phases, the treatment consisted in a multi-component intervention combining app-based GenAI (GPT-3.5) and peer-generated content shared through the app and Instagram, leveraging social-norm and feedback mechanisms, in a gamified framework. Using individual fixed-effects panel models, we assess the short-term effects of the multi-component treatment on key psychological determinants and behavioral intention to reduce holiday air travel. The intention-to-treat analysis provides no evidence of average treatment effects on either psychological outcomes or flying intentions. However, higher levels of engagement with the intervention are associated with lower intentions to fly for holiday purposes, and with a smaller reduction in perceived descriptive norms around holiday flying. While these findings should be interpreted cautiously given differential attrition (39% in the treatment group vs. 18% in the control group) and the study's limited power to detect small-to-moderate effects, they suggest that sustained engagement with interventions combining GenAI and peer-to-peer content sharing via social media may have the potential to influence the psychological antecedents and stated intentions related to holiday air travel.