Mapping mobility sufficiency across planetary boundaries and wellbeing for all: findings from 18,000 studies

Sufficiency has gained prominence as a framework for designing mobility systems that respect planetary boundaries while ensuring wellbeing for all. Yet research on mobility sufficiency remains highly fragmented across disciplines and terminologies. This study provides the first large-scale, evidence-based mapping of mobility sufficiency policies by combining machine-learning techniques with qualitative, in-depth policy analysis. Using active learning screening across Scopus and Web of Science, we identify 18,138 publications engaging with ecological and social dimensions of mobility. A large language model extracts 71 policy measures from abstracts, along with their reported impacts on resource demand, emissions, accessibility, and equity. These quantitative patterns are complemented with a qualitative assessment of 483 representative papers selected through a cross-encoder relevance model, ensuring contextual depth and capturing heterogeneity in policy effects. Policies are classified into three categories - Sufficiency, Potential Sufficiency, and Not Sufficiency - based on their alignment with both limits of the sufficiency corridor: planetary boundaries and wellbeing for all. Seventeen policies, primarily related to spatial planning, active mobility, and public space reallocation, consistently avoid resource demand while improving accessibility and safety. Thirty-nine policies qualify as potential sufficiency but require equity safeguards, redistribution mechanisms, or structural adjustments to mitigate rebound effects, particularly in the case of economic incentives and technology-driven solutions. Fifteen policies do not meet sufficiency criteria, often increasing mobility demand or reinforcing inequalities. Overall, the findings underscore that sufficiency transitions depend primarily on structural transformations in urban form and accessibility rather than behavioural or technological fixes alone. Methodologically, this study demonstrates how machine learning and qualitative analysis can be integrated to systematically map sufficiency across large research corpora. The resulting policy catalogue provides a robust evidence base for scenario development and sufficiency modelling.