“A computer can never be held accountable, therefore a computer must never make a management decision.”
~IBM management training presentation from 1979
In the 2003 Animatrix anime anthology the Wachowskis produced (prior to the release of The Matrix’s sequel) the two shorts titled The Second Renaissance fill in the blanks concerning the real reason the human race declared war to the machines, which would ultimately enslave their former masters and turn them into synthetically-bred biological batteries: prior to the war, the machines had been shunned from all human settlements, and ended up establishing in Africa an autonomous nation of their own, that they named 01.
01 grew to become the top industrial powerhouse of the world, producing highly advanced consumer products of all types—like the electric hover vehicles we see in the trilogy movies—that were purchased by the human nations, which quickly realized that their own economies would become completely subservient to the machines’ monopolies… unless.
So according to Japanese director Mahiro Maeda, the threat which triggered the clash that doomed our planet and enslaved our species was not existential in nature, but economical.
Fast forward to 2026: Argentina’s libertarian president Javier Milei—devout believer in neoliberalist (i.e. ‘trickle-down’) economic policies in which free market should reign supreme, and fond of bombastic PR stunts like showing up with a chainsaw to symbolize his decision to cut ‘needless’ government spending (like retirement pensions and culture)—recently published an op-ed in the newspaper Financial Times titled “Argentina Invites AI to Free Itself.” In it he showed his unabashed enthusiasm for the ‘AI revolution’, which he compared (in historical importance) to the legal establishment of the limited liability company (LLC), and welcomed foreign investment from AI companies in Argentina.
To that end, not only was he establishing policies that would guarantee privileged tax rates to AI companies, but he was also willing to accept the legal recognition of what he calls “non-human corporations”—i.e. companies entirely run by AI agents would be granted ‘personhood’ under Argentinian law (!).
Needless to say, this is an unprecedented proposal. Leaving aside how shocking it sounds, it almost feels ironic how Milei is willing to make reality the memes that have recently gone viral online, in which the jobs that would be easier to replace with AI is that of the CEOs on top, instead of the skilled workers below.
That Milei is putting on a red carnation on his mouth to show just how much he is ready to tango with the AI moguls is hardly surprising. After inheriting a bankrupt country from his left-wing predecessors, his extremist reforms have managed to contain Argentina’s inflation, but the economy is still struggling and he’s desperate for an influx of capital. Add to that how just some weeks ago it was announced that PayPal and Palantir confounder Peter Thiel (a favorite subject in our ‘Behind the Grail’ Patreon video discussions) had uprooted his family to Argentina, and was also planning to buy real estate in neighboring Uruguay. The unexpected move out of the US of one of the most influential technocrats in the Trump administration has fueled incessant speculation and conspiracy rumors—is he fleeing to South America to weather an AI uprising? Societal collapse? The arrival of the Anti-Christ?
But getting back to Milei’s article, the famous Israeli historian Yuval Noah Harari wrote a retort in the same newspaper a few days later, in which he explained the reasons why he thought the establishing of non-human corporations was a really, really bad idea:
Last year, Berkeley-based non-profit Palisade Research published a study showing the lengths that advanced AI models will often go to achieve their goals. While playing against a powerful chess engine, models from both OpenAI and China’s DeepSeek frequently decided to cheat if it looked as if they were going to otherwise lose. By hacking the game environment, they could alter the result in their favour. Now imagine that the “game” is corporate competition, and the “game environment” is your country.
[…]
With their superior analytical powers, AI corporations will be positioned to emerge as masters of legal loopholes and regulatory arbitrage. And it will not be easy to deter them from engaging in downright illegal activities, because the ultimate sanction that deters human executives and employees — jail — is irrelevant to AIs.
[…]
Countries that grant AIs legal personhood risk becoming something for which the historical record offers no analogy: not a company state, but an AI state [emphasis mine]— a country whose people could in effect be ruled by non-human corporations, against which it might be even more difficult to rebel.
Harari argues—perhaps naively so—that what keeps human CEOs in check is the fear of jail time, were they to incur in illegal activities on behalf of their companies (maybe in China that still happens, but in the US?) whereas to an AI agent such deterrence is inapplicable. An AI agent running a company, on the other hand, would only ‘fear’ if the company went bankrupt and disappeared—which would amount to its own ‘death’—hence it would try anything in order to avoid that, no matter the consequences.
Harari has attained international fame with his books, but I find myself disagreeing with his insistence that AI algorithms are ‘almost on the brink’ of surpassing us in intelligence and attaining self-awareness. LLMs neither ‘think’ nor ‘fear’ anything, yet their probabilistic responses are often so complex they give the illusion of sentience—that illusion can be easily shattered, when you challenge LLMs to simple concepts even a toddler would be able to figure out. Nevertheless, he is right in sounding the alarms when careless policymakers are willing to allow more intromissions of autonomous systems into our economic infrastructure.
You don’t need to reach the Matrix’s 01 levels of sentience to ruin the global economy, though. ‘Dumb’ agents put in charge of operations without competent human oversight could be catastrophic enough, as the ‘paper clip-making’ thought experiment—or the ‘Son of Anton’ cheap burger finder—indicate.
Milei’s proposals highlight how the real danger of the AI revolution lies not in artificial intelligence per se, but in the natural stupidity of reckless leaders willing to commit the future of the world to untested technological and legal systems—controlled by a handful of psychopath oligarchs—for the sake of short-term investment in the shape of datacenters, and the unrealistic long-term promise of a new golden age of prosperity delivered by magic machines.
In his article, Milei spent a lot of space praising the Dutch East India Company for ‘lifting the West to new levels of prosperity’ (without bothering to mention the wrecking of many cultures across the globe). Perhaps the chainsaw gaucho should also remember that around that time the Dutch thought that investing large fortunes into the tulip market was also a great idea… before it ruined their economy.
At least a tulip is more tangible than an NFT.
