Mitigating subjectivity and bias in AI development indices: A robust approach to redefining country rankings

Betania Silva Carneiro Campello, Guilherme Dean Pelegrina, Renata Pelissari, Ricardo Suyama, Leonardo Tomazeli Duarte

Abstract: Highlights•Methodological analysis using Choquet integral and SMAA enhances AI index robustness.•We combine SMAA and the Condorcet method to obtain more robust AI country rankings.•Interactions among AI dimensions are considered, mitigating bias in rankings.•SMAA assesses weighting influence on AI rankings, ensuring robustness.•Robust AI rankings derived without deterministic weight definition.