With AI systems being deployed across diverse societal contexts, the AI alignment problem has become critical. Ensuring that AI systems act in accordance with human values, particularly in complex multi-agent environments, remains a significant challenge. Conventional approaches—typically employing uniform value representations and consequentialist methods—often fail to capture the inherent variability in how values influence individual agents’ decision-making. In this work, we address this gap by extending the Computational Transcendence (CT) framework to integrate an agent’s sense of self into its decision-making process. Our approach embeds values within an agent’s identity, thereby facilitating adaptive, individualized behavior that accounts for external social influences, such as conformity. We demonstrate our model in a multi-agent simulation of transit decision-making, where agents choose between public and private transportation based on their value-driven identities. Our results suggest that using identity as a basis for value alignment offers a promising pathway for capturing human-like decision-making in AI systems.