Milei's

The Argentine government of Javier Milei has presented the so-called "Social Digital Twin," an artificial intelligence tool that, according to the Ministry of Human Capital, would allow for data integration, scenario simulation, anticipation of social problems, and more precise design of public policies.
The announcement was presented as a "paradigm shift" in social management, but it generated strong controversy due to a lack of details: what data will be used, who will develop the system, under what controls it will operate, what guarantees citizens will have, and what limits will exist regarding the political, commercial, or police use of that information.
To understand the debate, it's best to start with the basics. A digital twin is a virtual copy of something real, built with data, that serves to simulate behaviors and predict what might happen. The concept originated at NASA over 20 years ago to test machines, processes, and spacecraft without damaging them. The problem arises when this logic is applied to society, because it's no longer a turbine being simulated, but a population; it's not an assembly line being modeled, but life trajectories, consumption patterns, employment, health, education, poverty, or social conflict.
Milei's system would be a Social Digital Twin, a supposed artificial intelligence system capable of compiling data on health, education, work, income, social assistance, and living conditions to "predict" problems and "decide" public policies. It has even been alleged that it could process more than 30,000 data points per person, including demographic, educational, employment, social, and economic records.
An artificial intelligence tool applied to social policy doesn't operate in a vacuum. It functions within a specific state, with a specific government, and with a specific economic orientation. In Milei's Argentina, marked by austerity measures, the dismantling of public policies, and subordination to private capital, it's highly likely that AI won't be used to expand rights, but rather to classify the poor and anticipate conflicts.
Furthermore, the Argentine debate was ignited by the shadow of Palantir and its founder, Peter Thiel. Palantir didn't invent digital twins, but its own technology is presented as a way to integrate enormous volumes of data to build an operational representation of complex organizations. And it's not just any company: it's deeply linked to the US military, police, and intelligence apparatus.
The risks are not hypothetical. In the Netherlands, the tax agency used an algorithm to detect fraud in childcare benefits. The system ended up flagging low-income families with dual citizenship as suspicious. The result was devastating: some 26,000 families falsely accused, insurmountable debts, homes destroyed, and thousands of children separated from their families. The scandal led to the resignation of the Dutch government in 2021, and Amnesty International denounced it under the stark title: "Xenophobic Machines."
In Australia, something similar happened with Robodebt, an automated system that generated debts against social welfare recipients by cross-referencing income data. It falsely claimed approximately $1.76 billion from 526,000 people. The courts declared it illegal, and an investigative commission described it as a cruel mechanism, also linked to serious human harm.
The lesson is that when the algorithm is applied to the poor, the technical error becomes social punishment. What is presented as efficiency can end up being automated persecution.