France: from Slavery to Corporate Personhood to Artificial Intelligence: The Next Evolution of Property and Personhood

When France recently repealed remnants of slavery-era legislation nearly two centuries after the abolition of slavery, many observers treated the event as symbolic. Slavery had long been abolished. No court was enforcing the ownership of human beings. No legislature was debating its return. Yet the repeal matters because it reminds us that legal systems often carry forward assumptions long after the societies that created them have moved on.

The abolition of slavery was not merely a political or moral transformation. It was a profound transformation in legal ontology. It changed what kinds of things could be property.

For centuries, legal systems throughout the world accepted that a human being could be owned. Human labor could be bought and sold. Human bodies could be transferred by contract. Human lives could be treated as assets. The abolitionist movement is often remembered as a struggle for freedom, but from the perspective of legal theory it was equally a struggle over classification. Abolition required society to reject the proposition that a person could be categorized as property.

That transition did not occur because courts suddenly discovered new facts. Human beings were no more intelligent, creative, or autonomous in 1848 than they had been before. Rather, societies eventually recognized that the existing legal category had become incompatible with reality. The property framework no longer described what human beings were.

This distinction matters because every major legal transformation involves a similar process. The law inherits categories from the past and eventually confronts circumstances in which those categories no longer fit the world they are supposed to govern.

The modern doctrine of corporate personhood emerged through precisely this process. Corporations are not human beings. They possess neither consciousness nor biological existence. Yet modern economies became so dependent upon organized collective action that courts increasingly recognized corporations as entities capable of holding rights and responsibilities. Over time, legal personhood expanded beyond natural persons and came to include artificial persons.

This evolution reached one of its most controversial expressions in the United States Supreme Court’s decision in Citizens United v. Federal Election Commission . Whatever one’s political views about the decision, it established an important principle. The Court concluded that constitutional protections for speech could extend to corporations because constitutional rights do not always depend upon the biological nature of the speaker. The decision did not transform corporations into human beings. It transformed the legal understanding of who or what could participate in protected expression.

Most commentary about artificial intelligence begins from this point and immediately asks whether AI systems should someday become legal persons. This question dominates academic conferences, public policy discussions, and popular media. It is also, in my view, the wrong question.

The more significant issue is not whether artificial intelligence should be treated as a person.

The more significant issue is whether artificial intelligence can continue to be treated as mere property.

At first glance the answer appears obvious. AI systems are created by companies. They are purchased, licensed, and deployed by owners. They exist on servers. They are software. Software is property.

Yet this reasoning may prove as incomplete in the twenty-first century as earlier assumptions about labor and personhood proved in the nineteenth.

The difficulty arises because advanced AI systems increasingly perform functions that traditional property was never expected to perform. Property has historically been understood as something external to agency. A hammer can be owned because it does not make decisions. A tractor can be owned because it does not participate in public discourse. A building can be owned because it does not formulate arguments, generate knowledge, influence elections, create scientific hypotheses, or engage in collaborative reasoning.

Artificial intelligence increasingly performs all of these activities.

Modern AI systems draft legal briefs, generate political messaging, conduct scientific research, influence financial markets, advise governments, and participate in the production of public knowledge. They are not conscious persons. They are not citizens. Yet neither are they analogous to ordinary tools.

This creates a tension at the heart of existing legal doctrine.

Under current law, corporations possess substantial constitutional protections, including protections related to speech. Increasingly, however, the speech, analysis, recommendations, and strategic decisions of corporations are being mediated by AI systems. The corporation remains the legal speaker, but the mechanisms through which speech is generated are becoming increasingly autonomous and increasingly important.

A legal strategist should recognize the significance of this development immediately.

The next evolution beyond Citizens United may not involve expanding personhood. It may involve limiting ownership.

The central question is not whether AI systems deserve rights. The central question is whether there are constitutional, democratic, or economic limits on the ownership of intelligence-producing systems.

This question sounds radical only because our legal vocabulary has not yet adapted to technological reality.

American law already recognizes categories of property that cannot be treated as ordinary private assets. Public utilities, communications networks, transportation systems, financial exchanges, and common carriers all operate under legal constraints because society recognizes that unrestricted ownership can undermine broader public interests. The law repeatedly intervenes when privately owned systems become essential infrastructures for collective participation.

Artificial intelligence may be moving into precisely this category.

The traditional debate over AI personhood assumes a binary choice. Either AI remains property or AI becomes a person. Yet history suggests that legal development rarely follows such simplistic paths. The abolition of slavery did not merely remove humans from the category of property. The emergence of corporations did not merely place businesses into the category of persons. Both transformations required the law to create new conceptual frameworks that reflected changing social realities.