TECHNO-PATERNALISM IN POLITICAL RULEMAKING: AI OVERSIGHT VERSUS LEGISLATIVE SOVEREIGNTY

Abstract

AI's integration into political rulemaking introduces techno-paternalism, in which automated systems preempt legislative judgment in the name of efficiency, echoing classic tensions between harm prevention and individual autonomy. This study synthesizes fragmented literature, constructs a paternalism-sovereignty framework, and examines U.S and EU cases to diagnose erosion risks and design safeguards. Qualitative comparative case analysis of policy documents, GAO audits (2025), EU trilogies reports, and metrics (30% AI-influenced rules) tests propositions via process tracing and a 2x2 matrix. U.S. eCFR pilots exhibit hard paternalism, compressing 40% of debate cycles; EU AI Act sandboxes achieve proportionality via 18% vetoes, confirming efficiency-sovereignty trade-offs. Hybrid model 70% AI advisory, 30% legislative vetoes with XAI audits recommends adoption through national legislation and OECD standards, balancing legitimacy and speed gains.