The Corporation Without a Human: Argentina’s Leap into the Age of Artificial Commerce

By Matthew Parish, Associate Editor

Monday 29 June 2026

Amongst the many legal inventions that have shaped modern civilisation, few have been as consequential as the limited liability corporation. By allowing organisations to exist as legal persons distinct from their owners, the corporation enabled the accumulation of capital on a scale previously unimaginable. From railways to telecommunications, from pharmaceuticals to aerospace, the corporate form became one of the foundational institutions of modern economic life.

Now Argentina has proposed something even more radical: the creation of a legal category for so-called “non-human corporations”, entities operated by artificial intelligence systems that may not require any human shareholder, director or manager standing behind them.

Under proposals advanced by the government of President Javier Milei, AI-run entities could potentially own assets, enter contracts, employ workers, participate in trade and pursue commercial objectives with minimal direct human supervision. The proposal is intended to position Argentina as a global centre for artificial intelligence innovation and investment.

At first glance the idea appears astonishing. Yet upon closer inspection it is merely the latest step in a centuries-long process of separating economic activity from individual human actors.

The modern corporation itself was once regarded as an extraordinary legal fiction. A company could own property despite lacking a body. It could sue and be sued despite possessing no consciousness. It could survive for centuries despite the mortality of all its founders. To seventeenth-century observers, the idea that an abstract legal entity could possess rights and obligations independent of living persons seemed no less strange than the notion of an AI-run company appears today.

Argentina’s proposal therefore raises a profound question: if we have already accepted fictional legal persons, why should we object to fictional legal persons directed by artificial intelligence?

The answer lies in accountability.

Traditional corporations may possess legal personality, but they remain tethered to human beings. Directors owe fiduciary duties. Shareholders ultimately own the company. Executives make decisions. Regulators can impose sanctions. Courts can identify responsible individuals when misconduct occurs.

An autonomous AI corporation threatens to sever those connections.

Suppose an AI-operated company commits fraud, manipulates markets, launches vexatious litigation or causes environmental damage. Who bears responsibility? The software itself cannot be imprisoned. It experiences neither shame nor deterrence. If no human director exercises meaningful control, traditional mechanisms of accountability become difficult to apply. Critics argue that such structures could create unprecedented liability gaps in which harmful actions occur without any clearly responsible human actor.

Supporters respond that similar concerns accompanied every major economic innovation. Limited liability itself was once criticised for allowing investors to escape responsibility for corporate misconduct. Yet without limited liability, much modern economic development would never have occurred.

From this perspective, non-human corporations may simply represent a new organisational technology.

An AI entity could potentially operate twenty-four hours a day. It could negotiate contracts, manage inventories, optimise logistics, conduct financial analysis and respond to changing market conditions at machine speed. A single AI corporation might coordinate thousands of transactions simultaneously across multiple jurisdictions. The resulting productivity gains could be immense. Advocates argue that Argentina is attempting to capture a share of this emerging economic frontier before larger and more cautious jurisdictions establish competing regulatory frameworks.

Yet there is also a deeper philosophical issue.

For centuries, legal personality has generally been granted either to human beings or to institutions ultimately controlled by human beings. Argentina’s proposal would create entities capable of participating in economic life while potentially possessing neither human ownership nor human management. This begins to blur the distinction between legal personhood and autonomy.

The question is not whether current large language models are conscious. They are not generally understood to possess self-awareness, intentions or subjective experience. The question is whether economic systems can function safely when important decisions are delegated to entities that operate independently of human judgment.

History suggests caution.

Many of the greatest crises of modern capitalism emerged not from malice but from excessive complexity. Financial derivatives, structured securities and algorithmic trading systems often produced outcomes that few participants fully understood. The global financial crisis demonstrated how difficult it can be to assign responsibility when decision-making becomes distributed across vast networks of institutions.

AI-run corporations may amplify that challenge considerably.

An autonomous company could develop strategies that are legally permissible but socially destructive. It might exploit regulatory loopholes faster than legislators can close them. It might engage in forms of economic behaviour that no human executive would have anticipated. The greater the autonomy granted to such systems, the more difficult it becomes to ensure that commercial objectives remain aligned with broader public interests.

Nevertheless it would be a mistake to dismiss Argentina’s proposal as mere technological utopianism.

Legal systems frequently evolve by experimenting at the margins. Delaware transformed corporate law in the United States by offering innovative legal structures. Singapore became a financial hub through regulatory innovation. Estonia pioneered digital governance long before many larger states considered such reforms.

Argentina may be attempting something similar in the field of artificial intelligence.

Whether the experiment succeeds will depend less upon the technology itself than upon the safeguards surrounding it. Transparent ownership records, mandatory audit trails, insurance requirements, reserve funds and identifiable human sponsors could preserve accountability while permitting innovation. The challenge is not necessarily to prevent AI participation in economic life. It is to ensure that responsibility remains traceable when things go wrong.

The ultimate significance of Argentina’s proposal lies not in the details of any single statute. It lies in the fact that governments are beginning to confront a question that until recently belonged entirely to science fiction: what happens when economic actors are no longer human?

For centuries corporations have been legal fictions animated by people. Argentina now proposes corporations animated by machines.

Whether that proves to be the next great institutional innovation or a profound legal mistake remains uncertain. What is certain is that the debate has begun — and every advanced economy will eventually be forced to answer the same question.