RELIGIOUS EQUITY IN PUBLIC-SECTOR ARTIFICIAL INTELLIGENCE: DEVELOPING A MULTI-FAITH AI IMPACT ASSESSMENT FRAMEWORK THROUGH A COMPARATIVE STUDY OF TEXAS AND CALIFORNIA

Abstract

Artificial intelligence is increasingly embedded in public administration, education, health, benefits delivery, public communication, procurement, and regulatory decision-making. Yet state AI-governance frameworks rarely identify religion as a distinct domain of algorithmic equity, even though public systems can misclassify religious identity, reproduce stereotypes, fabricate beliefs, disadvantage minority-faith practices, or amplify sectarian misinformation. This article develops the Multifaith Artificial Intelligence Impact Assessment Framework (MAIIAF), a seven-domain instrument for evaluating religious representation, disparate impact, hallucination, community harm, transparency, remedy, and institutional accountability. Using structured comparative policy analysis, the article examines Texas and California as contrasting but nationally influential state models. Texas provides centralized statutory governance architecture, an AI council, state-use notice requirements, and an innovation sandbox. California offers a more operationalized procurement and risk-assessment system, mandatory inventories, vendor disclosures, workforce training, and lifecycle controls. The comparison finds that neither state presently provides a religion-specific testing protocol, affected-community consultation requirement, or standardized remedy for faith-related AI harms. MAIIAF addresses this gap by translating constitutional, civil-rights, risk-management, and interfaith-ethics principles into an auditable lifecycle process. The proposed framework is designed for adaptation by state agencies, school systems, public universities, health administrators, and technology vendors throughout the United States. Its national value lies in providing a replicable method for protecting religious liberty and equal treatment while supporting responsible innovation.