As Europe builds a regulatory framework meant to bind artificial intelligence to fundamental rights and classify systems according to their social risk, Javier Milei has chosen a different path altogether. He is inviting global capital to develop AI in Argentina through the legal invention of the non-human corporation: a commercial entity run entirely by software, with no flesh-and-blood shareholders, yet protected by capitalism’s most powerful device (limited liability).
The proposal by the Argentine president and his minister Federico Sturzenegger, published in the Financial Times, reflects the unease spreading through Silicon Valley after rulings such as Sarcuni v bZx DAO. In that case, a US court held that investors in an algorithmic protocol could be made personally liable for damage caused by the code. Against that backdrop, the Argentine initiative offers something close to a legal sanctuary: a place where AI systems or robots may sign contracts, experiment and fail, while the capital behind them remains beyond the reach of ordinary claims.
What Milei is proposing, in effect, is to recast Argentina as a haven for algorithms, a free zone in which capital can move beyond the ethical and legal constraints demanded by bodies such as the European Union.
Marginal legal spaces
Seen from a wider angle, this is about more than investment policy. What is taking shape is a subtler and more troubling geopolitical design: the creation of marginal spaces within the international legal order, places where practices unacceptable to advanced democracies can be relocated, shielded and normalised.
The logic is not entirely new. One saw a version of it in El Salvador in 2025, when Nayib Bukele made the CECOT prison available so that the Trump administration could illegally deport hundreds of Venezuelan asylum seekers, sidestepping US judges and brushing aside the principle of non-refoulement. Just as El Salvador offered its territory to externalise arbitrary detention in the name of security, Argentina is now offering its legal system to externalise algorithmic irresponsibility in the name of innovation.
Civilisational surrender
This form of legal extraterritoriality speaks directly to what Pope Leo XIV, in his recent encyclical Magnifica humanitas, describes as technofascism. The point is not merely that technology may do harm. It is that a society may be asked to absorb the consequences of decisions made by a neural network without there being any human legal subject left standing to answer for them.
To extend the corporate shield to machines in the legal havens of the Southern Cone is to take a decisive step towards civilisational surrender. It means weakening human sovereignty over technology at the very moment when that sovereignty matters most. The open question is whether Europe will have the political and legislative strength to prevent these grey zones of code from spreading to the outer edges of its own legal order.
This article is republished from Futuribles. Here's the original article in Spanish and English.