Argentina's Senate has opened formal debate on a bill that would do something no legislature has yet enacted: give artificial intelligence systems the legal standing to own and run companies. The General Legislation Commission, chaired by Senator Nadia Márquez, held its first hearing on June 24, 2026, on an executive-branch proposal that repeals Law No. 19,550 — the General Corporations Law that has governed Argentine business since 1972 — and replaces it with a framework built for algorithmic and blockchain-native firms (Senado de la Nación).
A 54-Year-Old Statute Meets DAOs
Deregulation Minister Federico Sturzenegger, Inspector General of Justice Alejandro Ramírez, and Undersecretary Paula Taddei presented the bill around five pillars: greater contractual autonomy, digitized registration, modernized entity types, stronger administrator accountability, and technological updating (Argentina.gob.ar). Ramírez's framing was blunt: the 1972 law was written for an industrial economy that predates the internet, and "a gap exists between corporate law and business practice."
The headline provision creates two new categories: sociedades automatizadas ("automated societies"), which can operate without human employees, and Decentralized Autonomous Organizations, governed by smart contracts and token-based membership. Both would receive legal personhood and limited liability. Government officials have stressed the bill does not rewrite civil or criminal liability — it layers new corporate forms on top of existing law rather than replacing accountability with a blank check (Senado de la Nación). Milei has compared the move to the 1602 chartering of the Dutch East India Company, arguing that the joint-stock corporation and the machine together built modern prosperity (Infobae).
Harari's Case, Steelmanned
Historian Yuval Noah Harari's public warning deserves to be taken on its strongest terms, not dismissed as technophobia. His core argument is an accountability asymmetry: "Human executives can be punished with jail terms for crimes and misconduct; artificial intelligence systems cannot." A corporation run by an algorithm with no human directing officer removes the deterrent that keeps ordinary corporate law credible — the threat of personal criminal exposure for the people actually making decisions. Harari's sharper claim is structural, not just moral: legal personhood functions as "a master key" to contract, borrow, hold assets, and sue and be sued, and handing that key to autonomous systems risks turning a company-state into what he calls an "AI-state" (Buenos Aires Herald). This is not a fringe objection. Beneficial-ownership opacity and judgment-proof shell structures are already chronic problems in cross-border finance; automating the operator only compounds them. Opposition senator Martín Soria raised a version of this concern directly in committee, questioning how anti-money-laundering oversight would function for firms with no identifiable human staff (La Nación).
Where Milei's Rebuttal Holds — and Where It Doesn't
Milei's response is stronger than critics credit. Legal personhood was never meant to erase human accountability, only to route it: corporations already act as legal persons distinct from their owners, and liability is enforced through the corporation's assets, dissolution, and civil penalties as often as through jailing individuals. Milei's point that "imprisonment is not the only way to deter unlawful conduct" is correct as a matter of existing law — most corporate wrongdoing is already punished through fines and asset forfeiture, not executive incarceration. Sturzenegger's defense in committee — that there's "no reason to think it will be less secure," comparing AI-run governance to trusting a GPS system — understates the actual novelty, but the underlying claim that ambiguity invites more risk than clear statutory categories is sound. Formalizing automated societies gives regulators a defined entity to subpoena, audit, and dissolve, rather than leaving AI-run ventures to operate in the legal gray zone they already occupy informally today.
Where the bill is genuinely underspecified is beneficial-ownership disclosure and a bonding or insurance backstop for automated societies with no human staff to sanction. Argentina doesn't need to abandon this experiment to address Harari's strongest point — it needs the Senate's ongoing hearings, which the committee has recessed to continue with outside specialists, to attach mandatory human-identifiable ownership registries and minimum capital or insurance requirements before final passage. That would preserve the deterrent Harari is worried about losing without forcing Argentina to forfeit the first-mover advantage Sturzenegger is chasing.
The Stakes of Being First
No other jurisdiction has attempted anything this comprehensive. If Argentina gets the guardrails right, it becomes a genuine testbed for how corporate law absorbs autonomous systems — a live experiment other legislatures will study rather than merely debate in the abstract. If it doesn't, Harari's "AI-state" framing will look less like alarmism and more like an accurate forecast. The distinction won't be decided by the headline concept of "non-human corporations," which is legally coherent enough, but by the disclosure and liability details the Senate commission is still writing.