Should Artificial Intelligence (AI) systems have the vote? One approach to this question is to explore AI systems’ moral characteristics, asking whether they possess the qualities that entitle humans to equal political standing and, consequently, suffrage. Those are important questions. But they are very hard ones, resting on highly contestable premises about ethics, AI, and humanity.

We instead take a functionalist approach: On the current trajectory, AIs will soon be highly agentic and capable, while remaining difficult to align with human objectives. If so, law will be an essential tool for shaping such systems’ behavior to favor positive-sum cooperation with humans, rather than conflict.

This Essay applies that functionalist approach to perhaps the most foundational legal question in democratic societies: Who has the franchise? Building on prior work arguing for AI private rights in tort and contract, we analyze the case for giving AI the right to vote.

We canvas four instrumentalist theories of democracy and ask whether each supports extending political rights to AI. Specifically, we argue that AI suffrage may: (1) reduce powerful AI systems’ incentive to disempower humanity by credibly committing public policy to account for AIs’ preferences; (2) promote economic growth by forcing governments to invest more heavily in public goods; (3) improve the epistemics of policymaking by adding competent and diverse voters whose reasoning is transparent and queryable; and (4) improve the “felt legitimacy” of government, at least insofar as that concept tracks AI behavior rather than genuine emotion.

We then consider three powerful objections to AI suffrage: the problem of how to count AIs, the risk that AI may swamp human voting, and the risk that AIs or AI companies may seize disproportionate power. We propose concrete institutional responses, including “algorithmic corporations” (“A-corps”) as the unit of AI political identity, vote-share caps pegged to AI economic contribution, and “lottocratic” selection among AI voters. Throughout, we argue that the distributional costs of AI suffrage are likely outweighed by growth effects over time.