Milei’s “non-human corporations”: a radical bet on AI legal personhood and corporate autonomy
Argentine President Javier Milei’s proposal—floated in a Financial Times opinion piece—to recognize “non-human corporations” governed entirely by artificial intelligence is more than a provocative thought experiment. It is a direct attempt to re-engineer the country’s economic positioning through AI-first corporate law, pairing technological ambition with a deregulatory ideology that treats oversight as friction rather than protection.
The concept is intentionally framed with historical resonance. By invoking the Dutch East India Company, Milei signals a vision of Argentina as a modern “free port” for global commerce—this time built on algorithms rather than shipping routes. Yet the analogy carries a double edge: the VOC represented not only innovation in corporate structure and capital formation, but also an enduring legacy of extraction, coercion, and asymmetric power. That tension sits at the heart of the debate: whether Argentina is proposing a bold new institutional form for the AI economy, or inviting a replay of concentrated power—now automated and scaled.
Milei’s legislative pillars, as described, are stark in their simplicity: no AI-specific regulation, legal recognition of AI-managed entities (with optional human shareholders), and minimal corporate taxation. Taken together, they amount to a deliberate invitation for regulatory arbitrage, positioning Argentina as a jurisdiction where AI systems can operate with maximal freedom and minimal constraint.
From DAOs to AI-run firms: what changes when algorithms become the boardroom
The closest precedent is not traditional corporate governance but the world of Decentralized Autonomous Organizations (DAOs)—structures that attempted to encode governance into software and token incentives. Milei’s proposal effectively asks: what if the DAO ethos were transplanted into mainstream corporate law, without the friction of heavy compliance regimes?
The technological promise is easy to articulate. AI-managed corporations could, in theory, deliver:
- Faster decision cycles, optimizing pricing, procurement, and capital allocation in near real time
- Lower administrative overhead, reducing layers of management and internal bureaucracy
- Data-driven governance, where strategy is continuously updated based on market signals
But the risks are not merely theoretical. Fully autonomous decision loops can amplify problems that human governance often dampens:
- Algorithmic bias becomes a governance feature, not a bug, if no accountable human layer exists to challenge outcomes.
- Opaque feedback loops—such as self-reinforcing cost cuts, labor suppression, or aggressive market capture—can become “rational” under a narrow optimization function.
- Model brittleness emerges when systems trained on historical data face shocks: political instability, supply chain disruptions, or sudden regulatory shifts abroad.
The hardest question is not whether AI can run a company, but whether a society can assign responsibility when it does. Corporate law is built around fiduciary duty, intent, negligence, and accountability. An AI-run entity strains each concept: if an autonomous system causes harm, who is liable—the developer, the deployer, the shareholder (if any), the hosting jurisdiction, or the algorithmic entity itself?
Argentina’s fiscal and regulatory gambit: investment magnet or race-to-the-bottom dynamics
Milei’s plan is also a macroeconomic strategy. Argentina has long struggled with credibility, inflationary history, capital flight, and debt fragility. A “free port” approach aims to reverse the narrative: make Argentina a destination for foreign direct investment (FDI) in AI, compute infrastructure, and corporate domiciling.
A low-tax, low-regulation environment can attract early movers—particularly firms constrained by tighter regimes such as the EU AI Act or more fragmented U.S. compliance expectations. Yet the same design introduces structural vulnerabilities:
- Reputational risk: global enterprises may hesitate to base sensitive AI operations in a jurisdiction perceived as a haven for unaccountable automation.
- Regulatory unpredictability: if political winds shift or international pressure mounts, the legal environment could change abruptly—raising the cost of capital.
- Currency and sovereign risk feedback: inflows can support the peso and improve credit optics, but sudden outflows—triggered by controversy or sanctions risk—can intensify volatility.
Regionally, the proposal could pressure neighbors—Brazil, Chile, Colombia, and Mercosur partners—to loosen standards to remain competitive. That dynamic is familiar in tax competition; applied to AI governance, it risks fragmenting Latin America into a patchwork of policy shopping, where firms choose jurisdictions not for excellence but for permissiveness.
Accountability, inequality, and geopolitics: the real stress tests for AI corporate governance
The most consequential implications are ethical and geopolitical, because they determine whether the experiment becomes a sustainable model or a cautionary tale.
Accountability and legal liability sit at the center. A corporation is already a legal fiction; making it “non-human” in governance pushes that fiction further. Without clear frameworks for:
- tort liability and negligence standards,
- auditability and explainability expectations,
- enforceable compliance obligations, and
- insurance and indemnification mechanisms,
the market may price these entities as inherently hazardous—raising financing costs and limiting mainstream adoption. Conversely, a robust private ecosystem could emerge around them: specialized insurers, AI risk auditors, and legal engineering firms that translate algorithmic behavior into contractual commitments.
Inequality risks are equally salient. If AI-run corporations optimize relentlessly for profit, and if human shareholders are optional, wealth could concentrate around those who control:
- proprietary models,
- high-quality data assets, and
- compute infrastructure and cloud distribution.
That concentration may not remain domestic. Argentina could host the legal shells and operational autonomy while value accrues to foreign capital and platform owners—recreating a modern form of extraction, albeit through digital channels rather than commodities.
Finally, geopolitics and digital sovereignty loom. A permissive regime could make Argentina attractive to projects blocked elsewhere, including those navigating export controls, surveillance concerns, or restricted model deployment. That may increase Argentina’s strategic relevance—but it also raises the prospect of becoming a conduit for contested technology flows, inviting diplomatic pressure and complicating trade relationships.
Milei’s proposal is, at its core, an attempt to turn corporate law into a competitive technology instrument—treating AI governance as a market to be won. Whether Argentina becomes a genuine hub for responsible AI-enabled enterprise or a jurisdiction defined by unbounded algorithmic power will depend less on the rhetoric of freedom and more on the unglamorous mechanics: liability, auditability, enforceability, and the capacity to ensure that autonomy does not become impunity.
By
By
By
By
By
By
