Key takeaways
- Argentinian President Milei’s non-human corporations bill, for all its ambition, still requires human administrators, human legal representatives, and human compliance officers. And yet, the headlines and the legislation are telling very different stories.
- The evidence from companies like Ford, where over 350 veteran engineers had to be rehired after AI systems fell short, shows that removing humans from operations is considerably easier than removing them from the accountability chain.
- AI accountability cannot be automated. Someone still has to answer when things go wrong, and that, more than any legal framework, is why non-human corporations will remain, for now, a compelling idea rather than an operating reality.
I had to read it twice.
“Argentina moves to legalize non-human corporations run entirely by AI.” The words belonged to Javier Milei, Argentina’s right-wing libertarian president. I put my tea down, went back to the beginning of the sentence, and read it again. Not because I didn’t understand it, but because I wanted to make sure I had. Milei’s vision: a company with no humans required; no CEO, no board, no shareholders, just AI agents signing contracts, owning assets, making decisions, with a legal framework to match; Buenos Aires as the next great hub for global commerce.
The man himself, chainsaw in hand at conservative gatherings one year, rewriting corporate law the next. I’ll give him this: he knows how to make headlines.
In a Financial Times op-ed published in early June, Milei announced a congressional bill built on three pillars: keep AI unregulated, create a new legal category called the non-human corporation, and offer a low corporate tax rate to attract tech investment. The historical reference he chose to anchor it was the Dutch East India Company, founded in 1602, which he positioned as the legal innovation that made capitalism possible. Buenos Aires, he suggested, could do for AI what Amsterdam did for the age of sail.
It is a genuinely bold premise. And if you only read the headlines, it sounds revolutionary. If you, however, read the actual bill, it gets considerably more complicated.
What the non-human corporations bill doesn’t say
The “automated company” introduced in the proposed reform requires a human administrator to oversee operations. Directors who use AI for decision-making remain personally liable for the outcomes, as Article 102 of the draft law explicitly preserves that liability. The DAO structure, the closest the bill comes to full autonomy, still requires a human legal representative, a human promoter with unlimited liability, and a human compliance officer wherever anti-money-laundering rules apply.
Lawrence Cunningham, director of the Weinberg Center for Corporate Governance at the University of Delaware, was more measured: “We are not changing the world here so much as we are recognizing that you might run a business without any HR.”
The law has yet to pass Congress. According to government representatives, not a single company or investment commitment is attached to it. Not quite the Amsterdam of the AI age.
So, can companies thrive without humans?
What Milei has legislated beneath all that spectacle we know he loves, is something far more modest: a formal acknowledgment that AI can run day-to-day operations, while a human still sits at the end of the accountability chain. Which raises a real question, and one worth asking: can companies genuinely thrive without humans? Not just without human shareholders, but without human judgment, institutional memory, and the instinct that comes from experience.
The evidence keeps delivering the same answer.
Ford spent years replacing experienced engineers with AI quality systems. The automated tools were supposed to handle quality checks across production. They couldn’t. Over 350 veteran engineers had to be rehired, brought back specifically because their decades of accumulated knowledge had never been transferred before they left. Charles Poon, Ford’s VP of vehicle hardware engineering, admitted it publicly: “Mistakenly, we thought that by just introducing artificial intelligence and ingesting the design requirements that we had, that would produce a high-quality product.” What emerged was actual human AI collaboration: the engineers training the systems, the systems flagging the problems, the humans making the final call. The rehiring effort contributed to millions of dollars in cost savings and helped Ford top the JD Power Initial Quality Study for the first time since 2010.
Ford is not an outlier. The fantasy of AI replacing humans entirely is not surviving contact with reality. According to a Forrester Research report, 55% of employers who restructured their workforce through AI-driven layoffs in 2025 and 2026 now regret the decision. Likewise, a survey by Careerminds found that nearly a third of companies that conducted AI layoffs had already rehired for 25-50% of the roles they cut, and over half said the rehiring crisis hit them within six months. The roles returning fastest are mid-level managers, customer success directors, quality assurance specialists, in short, the people who hold institutional memory, navigate unspoken dynamics, and repair relationships when an algorithm gets it wrong.
A great headline, a flawed premise
This is what Milei’s bill, for all its ambition, quietly admits. You can remove the human from the org chart. You cannot remove the human from the accountability structure. AI accountability cannot be coded into an algorithm. Someone still has to answer when things go wrong. And things will go wrong, trust me, because they always do, and because the deterrents that keep human executives in line mean precisely nothing to a piece of software, however perfect and “smart” it may be.
Milei will probably get his law passed. Argentina needs the investment, the positioning, the signal to Silicon Valley that it is open for business. And, if we are completely honest, it also needs some kind of legal framework to deal with the reality of AI. In time, the bill may well attract capital, registrations, early movers looking for a permissive jurisdiction ahead of tighter regimes elsewhere. On those terms, it may succeed.
But the non-human corporations, as a genuine operating reality, the companies that run, compete, and thrive without human judgment at their core, will remain just that: a great headline. Reality, as Ford and hundreds of other companies are currently finding out the hard way, keeps finding its way back to humans.
