Configurando la ciudadanía sostenible asistida por IA: Un estudio fsQCA del modelo CAITIZEN

Resumen

Context. Artificial intelligence increasingly mediates learning, decision-making, creativity, and citizenship formation in higher education, requiring critical, ethical, fair, collaborative, and reflective capabilities for sustainable development and responsible human–AI interaction. Problem. Previous CAITIZEN studies established its qualitative architecture and PLS-SEM relationships, but net effects do not reveal sufficient configurations or asymmetry. Which configurations of critical AI literacy (CAIL), ethical responsibility (EARB), fairness and data justice (AFDJ), human–AI collaboration (HAIC), and metacognitive prompting (MTPP) are sufficient for high versus low CAITIZEN?

Purpose. This study compares these configurations through a multidisciplinary sustainable- innovation perspective informed by the OECD Oslo Manual, primarily aligned with SDG 4 and complementarily with SDG 9.
Methodology. An fsQCA study used data collected from July to December 2025 in Jalisco, Mexico, from 511 higher-education students. Five ten-item conditions and the ten-item CAITIZEN outcome were calibrated using P5, 5.00, and P95. Necessity, sufficiency, coverage, consistency, core/peripheral roles, asymmetry, and robustness were assessed.

Theoretical and Practical Findings. No condition was necessary. Five configurations were sufficient for high CAITIZEN (coverage=.866; consistency=.917), while low CAITIZEN followed a distinct configuration (coverage=.600; consistency=.924). Scientia extends CAITIZEN toward a configurational theory of sustainable AI-assisted citizenship; Praxis provides alternative capability architectures for responsible AI education and sustainable innovation primarily aligned with SDG 4 and complementarily with SDG 9.

Originality. CAITIZEN integrates a multidisciplinary sustainable-innovation perspective with configurational analysis, revealing multiple pathways through combinations of human and technological capabilities.
Conclusions and Limitations. CAITIZEN is configurational, conjunctural, equifinal, and asymmetric. Its cross-sectional, non-probabilistic sample limits generalization and temporal inference, requiring replication to assess stability.