Explainable AI-driven predictive governance of Gig workers in Indonesia: Employment characteristics, social security accessibility, and policy recommendations

This study investigates the employment characteristics associated with BPJS exclusion among gig workers in Indonesia through an Explainable AI–driven predictive governance framework. Using survey data from 1,047 gig workers across Jakarta, West Java, Banten, and Yogyakarta, the study examines how demographic, occupational, and platform-related factors relate to social security accessibility within Indonesia's expanding digital labor economy. Logistic Regression, Random Forest, and XGBoost models were employed to classify BPJS participation status, followed by policy simulation scenarios to explore potential intervention outcomes. The findings indicate that BPJS exclusion was comparatively higher among ride-hailing and food/courier workers, workers with unclear contractual arrangements, and respondents highly dependent on platform income. Online freelancers and digital marketplace workers demonstrated relatively lower exclusion rates. Among the predictive models, Logistic Regression showed slightly stronger overall performance, although all models demonstrated moderate classification capability. The models were generally more effective in identifying excluded workers than active participants, suggesting that exclusion patterns were more consistently represented within the dataset. Policy simulation results further showed relatively limited differences across intervention scenarios, although platform co-contribution mechanisms produced slightly more stable participation outcomes than digital enrollment facilitation and premium subsidy expansion. The study suggests that social security exclusion among gig workers is shaped by heterogeneous employment conditions and platform structures. While the predictive models remain exploratory, the findings demonstrate the potential contribution of explainable analytical approaches for supporting more evidence-informed discussions on inclusive social protection governance in Indonesia's evolving gig economy.