Agentic Inequality

Matthew Sharp, Omer Bilgin, Iason Gabriel, and Lewis Hammond ask how autonomous delegates could redistribute power and opportunity. Their framework separates agent availability, quality, and quantity.^[1]

Method and contribution

The conceptual analysis maps these dimensions to economic and political mechanisms: scalable delegation, agent competition, infrastructure access, and platform dependence. Equal access to a model need not mean equal access to a functional agent; user resources can also affect benefits. The paper considers both concentration and equalizing possibilities.^[1]

Alignment relevance and limits

Agentic Inequality makes Social Alignment a distributional question, beyond obedience to an owner. The authors propose measurement, participatory fairness, interoperability, and public-service research. These are a research agenda, not demonstrated policy effects or a prediction that inequality inevitably rises. Evidence from assistive tools may not transfer to autonomous workflows.

See Aligned with Whom? Direct and Social Goals for AI Systems.

Historical context

Sharp, Bilgin, Gabriel, and Hammond first submitted Agentic Inequality to arXiv on October 19, 2025. This date records the first submission, not a later revision or institutional release.^[1]

See Agentic Inequality.