Argentina Launched an AI to Predict the Future. It Couldn’t Predict a Typo

Argentina Ministry of Human Capital has a bold claim: it can predict the future of social policy using artificial intelligence. President Javier Milei announced the “Gemelo Digital Social” initiative (which roughly translates to “Social Digital Twin”) via X on Friday, calling it “a paradigm change in social policy.”

He closed the announcement with “MAGA. VLLC!” – a nod to Trump’s slogan next to his – lest anyone miss the political branding.

The system, a ‘social digital twin’, is designed as a virtual, dynamic replica of Argentine society. It collects data from multiple government and private sources and then uses AI to simulate scenarios, anticipate consequences, and optimize policy decisions in real time.

The stated goal: To move Argentina from a “reactive state” – one that responds to social problems after the fact – to a “predictive state” that can model poverty, track the effects of subsidies and chart human capital development from childhood to adulthood.

Digital twins are established technology. They have been used for years in engineering, urban planning and infrastructure, simulating how a bridge holds up under load, or how traffic flows before a road is built. The Argentine government claims this would be the first time the concept has been applied to social policy on a national scale.

The system would collect data, identify patterns, project scenarios and convert social experiences into what the ministry calls “public intelligence.” In practice: a centralized database drawn from government agencies and the private sector – healthcare, income, education, consumption – fed by an AI model that tells policymakers what will happen. Think of it as a weather forecast for poverty.

It is not an unprecedented idea in government. Declutter reported in April 2025 that the UK Ministry of Justice was secretly building an AI system to predict who might commit murder, by collecting mental health data, addiction histories and self-harm reports from more than 100,000 people. That program immediately drew comparisons to Phillip K. Dick’s novella and film “Minority Report” and set off a civil liberties firestorm.

The Argentine goal is softer – social optimization rather than crime prediction – but the underlying architecture is similar: collect enough personal data and let an algorithm tell you what happens next.

The Internet responds

The vision was futuristic. The execution was not.

The promotional video released to announce the Gemelo Digital was full of mistakes that immediately led to ridicule. At 0:35 there was an image saying “MULTIPLES FUENTES” – missing the obligatory accent on the esdrújula-múltiples. The bigger blunder appeared at 0:54: a full-screen declaration that the system was the “PRIMER SISTEMA QUE AYUDA PREDICIR EL FUTURO” – dropping the preposition “a” before the verb (which makes the whole thing sound weird in Spanish) and misspelling “predecir” as “predicir”.

The digital twin system that promises to predict the future couldn’t predict a typo.

“No predijo los errors de ortografía,” user @pablomen0 joked on X: “It didn’t predict the spelling mistakes.”

Developer and tech commentator Maximiliano Firtman chronicled the full shame: “Grammar and spelling errors, a fake minister presenting holograms, Singaporean flags, Amazon AWS logo, a terrible speech. Unbelievable.”

It fits a pattern. Just a few weeks ago an official photo of Milei at his desk in Casa Rosada (the presidential palace) went viral for through the window behind him Casa Rosada appeared again. An AI-generated image of a president, in the palace, looking out towards the same palace. The presidency’s digital communications team has a recurring problem with unattended AI output.

The political response came quickly. Opposition Senator Agustín Rossi filed a formal information request demanding transparency on the program’s legal framework, data protection and civil rights guarantees. “The future cannot be surveillance of citizens,” Rossi said wrote on X. Milei’s government — which has faced repeated scrutiny over its relationships with tech companies since the Libra coin scandal — has not publicly addressed the governance issue.

Privacy experts went further. Mass aggregation of real data on Argentine citizens requires legally strict anonymization protocols. No such framework has been announced.

Analyst Julián Roô the care framed at the structural level: “Argentina will be the laboratory rat for analyzing how a society works when algorithms classify citizens based on risk, productivity or behavior. Starting today, Argentina will move from social policies based mainly on human decisions to automated predictive systems fueled by AI and big data.”

Political analyst Pablo Munoz Iturrieta wrote: “It sounds futuristic and efficient. The point is, this sounds like every authoritarian technocrat’s wet dream.”

Senator Rossi’s formal request for information remains pending. The Ministry of Human Capital did not elaborate on the video errors or the data governance framework under which the system would operate.