Territorial Governance and Technological Convergence: Toward a Methodological Framework for Social Innovation Based on Artificial Intelligence and the Multi-Helix Model from the Global South

Social Innovation (SI) has emerged as a strategic paradigm for addressing systemic challenges in highly uncertain environments. However, its practice still reveals epistemological fragmentation that risks reducing SI to welfare-oriented approaches. This article presents a critical and constructive analysis aimed at mitigating the “methodological myopia” that persists in social impact assessment. Through a systematic literature review and a qualitative case study in the Antofagasta Region (Chile), the article argues that the scientific validity of SI depends on longitudinal, multidimensional, and territorially grounded evaluative frameworks. The study examines the relationship between SI and Artificial Intelligence (AI) as a source of methodological rigor, improving traceability and auditability while supporting the scaling of interventions. In response to techno-utopian forms of determinism, the article proposes an ethical and participatory governance framework based on the Multi-Helix model, integrating academia, the public sector, private enterprise, and civil society in the co-creation of public value. The findings suggest that the institutionalization of SI through AI must move beyond procedural efficiency to foster structural transformation. In Antofagasta, this AI-supported certification architecture is already operational within the Regional Innovation Strategy (ERI) 2022–2028 through the executed FIC-R 2023 project on Social Innovation Certification. The Antofagasta experience is therefore presented as an illustrative case of territorial governance, offering transferable principles for other Global South contexts rather than a directly replicable model.