Argentina’s AI-driven businesses still rely on humans.

      The Argentine government aims to allow a company to operate without a human in charge and has proposed a bill to legalize this. The legislation introduces a new classification known as the non-human corporation, which would be managed by AI agents or robots capable of entering contracts and managing assets independently, with human shareholders optional. However, despite its initial framing, the bill relies on human involvement more than President Javier Milei implies, especially as governments globally are moving towards stricter AI regulations.

      The most independent model proposed, structured on a blockchain as a decentralized autonomous organization, still necessitates a human legal representative for any actions needing a signature, alongside a human promoter who would assume unlimited responsibility for the company's obligations during its formation, according to legal assessments of the proposal. Additionally, where anti-money laundering laws are relevant, a human compliance officer is also required. The government presented the draft in May, intending to replace the current general corporations law in Argentina, which has been in effect since 1972. Officials have positioned the reform on three foundations: ensuring AI remains unregulated, establishing the non-human corporation category, and offering a low corporate tax rate to attract tech investments to Buenos Aires.

      This proposal is distinct from Super RIGI, the government’s parallel initiative aimed at incentivizing large AI data centers, but both focus on the same investor demographic. The proposal is competitive; Milei’s administration aims for Argentina to become the preferred destination for AI enterprises that would prefer minimal regulatory oversight and the absence of a human board. However, whether this ambition can withstand the implications of the bill's liability clauses remains uncertain. Legal analysis has pointed out that even the purportedly autonomous structures maintain a "human floor," since a director overseeing or configuring an AI system is still responsible for its actions.

      The proposal has faced sharp criticism from historian Yuval Noah Harari, who contended that the absence of an accountable human in corporate decisions creates the very liability void that corporate law strives to avoid. Harari referenced Milei’s own analogy linking the plan to the Dutch East India Company, noting that this company’s most critical action was setting fire to the port of Jayakarta in 1619 and subsequently ruling the area as a private empire. He cautioned that Buenos Aires risks transforming into a “new Batavia” instead of becoming a financial center.

      Mustafa Suleyman, the CEO of Microsoft AI, echoed similar sentiments, arguing that AI agents should not be granted greater legal recognition than a laptop, emphasizing that developers should work against the idea that their systems possess quasi-personhood deserving of rights. Milei engaged in a lengthy response on social media, asserting that recognizing AI-operated entities with a defined legal status would enhance regulation, as it provides regulators with a specific structure to address rather than an unaccountable software operating outside established law.

      This perspective hinges on the reliability of the legal categories once AI-run entities scale up—a situation that has yet to be tested in any jurisdiction. Critics argue that the usual deterrents keeping human executives accountable, notably the potential for prosecution, have little effect on algorithms, and a promoter’s infinite liability may fall short when the entity undertakes decisions beyond human comprehension.

      It remains uncertain whether Argentina’s Congress will pass the bill as it stands, and whether other governments will follow suit. This discussion emerges as regulators in other regions are tightening, rather than loosening, regulations concerning autonomous systems. Advocacy groups pushing for the EU’s AI regulations call for increased human oversight of automated decisions, and companies like SAP are restructuring their executive oversight of AI, rather than eliminating it. Argentina’s bill represents a gamble that a more lenient approach will prove more successful first, with a person still accountable behind the proposal, no matter how it is marketed.