Beyond western paradigms: a cross-cultural review of social–emotional competence frameworks and the promise of AI-driven assessment

Background Social–emotional competence (SEC) is a critical yet culturally embedded construct. While foundational frameworks like collaborative for academic, social, and emotional learning (CASEL) and OECD provide influential models, their cross-cultural applicability and the validity of associated assessments are increasingly questioned, particularly in non-Western contexts. Concurrently, artificial intelligence (AI) presents novel opportunities for culturally responsive SEC measurement. Objective This systematic review synthesizes literature on the conceptualizations, theoretical frameworks, and measurement tools for adolescent SEC across diverse cultural settings, with a specific focus on the role of emerging digital technologies. Methods Following PRISMA 2020 guidelines, a comprehensive search of Web of Science and Google Scholar (2000–2025) was conducted. After AI-assisted triage and manual screening, 89 studies were included. A narrative synthesis was performed, employing thematic analysis to compare frameworks (e.g., CASEL, OECD, Chinese, Australian/New Zealand models) across theoretical foundations, operational dimensions, and measurement approaches. Results The analysis reveals that dominant SEC frameworks are cultural artifacts reflecting underlying individualist or collectivist values, leading to divergent prioritizations of competencies (e.g., autonomy vs. harmony). Achieving cross-cultural measurement invariance for standardized tools remains a significant challenge, necessitating strategies like anchoring vignettes and emic–etic integration. The review identifies a clear trajectory toward technology-enhanced assessment, highlighting the potential of multimodal AI analysis, generative AI for stimuli creation, virtual reality simulations, and large language models to enable more ecologically valid, behavioral, and culturally configurable evaluations. However, these technologies introduce risks of algorithmic bias and digital colonialism. Conclusion Advancing the field requires a pluralistic, dialogical approach that decentralizes Western models, invests in indigenous theory-building, and ethically harnesses technology. Future research must develop assessment methodologies that balance generalizability with deep cultural respect, leveraging AI as a tool for empowerment and context-rich insight rather than for imposing reductionist, cross-cultural rankings.