Ethical for Whom: A Systematic Literature Review of Multicultural Perspectives on Responsible AI Values for the Global Majority

Responsible AI principles like fairness, accountability, and transparency are built on Western institutional assumptions yet exported as universal standards. We systematically reviewed (N=41) studies from underrepresented regions to identify locally prioritized values and their tensions with dominant discourse. Three frictions emerged: epistemic (whose knowledge counts as ethical expertise), contextual (whether abstract principles survive local conditions), and infrastructural (how resource asymmetries constrain value realization). Our findings show Global Majority scholarship does not simply add new values to the list. They reconfigure ethics through lived experience by redefining the core logic: fairness shifts from individual parity to collective well-being; transparency shifts from technical disclosure to relational trust but bounded by protocol and custom; and accountability shifts from top-down audit to distributed responsibility shaped by labor conditions and data sovereignty. This study advocates for centering epistemic pluralism as a prerequisite for ethically legitimate AI governance rather than treating these frictions as implementation failures.