Algorithmic authority in Spain: relative deprivation, hierarchy, and support for AI rule

Artificial intelligence is increasingly used in public administration. Whether citizens would grant it authority over decisions normally entrusted to representative institutions is a separate political question. Using cross-sectional data from a quota-stratified Spanish survey, this article examines support for AI political authority. Respondents evaluated three propositions: that AI should make important parliamentary decisions, that it should oversee lawmaking, and that AI-based political decision-making would reduce concerns about corruption. We conceptualize endorsement of these propositions as bounded algorithmic exit: openness to relocating political judgment from representative institutions in specified domains. The analysis combines descriptive estimates and multivariable models with an ordinal latent-variable model estimated using WLSMV. In the attention-check-passing sample (N = 1,490), support was uncommon: 15.2% scored above the scale midpoint, and 5.4% scored at least 5 on all three items. Across full-distribution models, support was most consistently associated with greater relative deprivation, stronger social dominance orientation, more right-leaning ideology, and younger age. The corruption proposition attracted more agreement than either proposition assigning AI direct political authority. The study distinguishes AI as an administrative tool from AI as a political authority and connects support for AI rule to technocracy, social hierarchy, and democratic accountability.