Behavioural Science and AI-Driven Interventions in Energy Conservation: A Systematic Literature Review

This systematic literature review synthesises findings from 45 peer-reviewed studies (2016–January 2026) on behavioural interventions for energy conservation. Key results show that social comparison nudges lower electricity use by 1.5% to 2.9%; about 52% of savings persist post-intervention via technology. Intervention efficacy is 2–4 times higher among liberals versus conservatives and is context-dependent. AI-personalised interventions outperform static nudges, reducing decay effects by up to half. Boosts foster sustained performance, with 26.6% greater long-term effects than nudges. Notable challenges include moral licensing (5–8% rebound), boomerang effects among low users, and publication bias that inflates impact estimates. There are gaps in cross-cultural validation, long-term persistence, transparency, trust, and integration in energy planning. This review provides actionable policy recommendations for designing ethical and effective behavioural programs that support technology-driven climate action.