By José Luis Sampietro*/Latinoamérica21
On the threshold of a new technological era, Latin America observes the advance of artificial intelligence with a certain mix of hope and fear. In official speeches and business forums it is repeated with enthusiasm that the region must get on the innovation train, but on the streets, in the workshops and in the classrooms, the question is different: Who really designs the future of our jobs?
If Latin American history teaches anything, it is that technological revolutions, when they arrive without politics and without equity, tend to widen the distance between those who decide and those who obey. Recent data from the International Labor Organization are eloquent. Between 26 percent and 38 percent of jobs in Latin America and the Caribbean could be affected by the emergence of generative artificial intelligence. Of that universe, between 8 percent and 14 percent would experience productivity improvements thanks to the use of automation and analysis tools, while between two percent and five percent are at risk of disappearing completely. Behind these percentages are real people: administrative workers, technicians, salespeople, teachers or independent professionals.
Artificial intelligence silently reconfigures tasks, hierarchies and opportunities. And in a region marked by structural inequalities, the impact will not be uniform. While large urban companies can invest in training, connectivity and digital transformation, millions of informal workers, young people without access to technical education or rural communities without a stable connection are left out of the map of progress. The World Bank estimates that between 30 percent and 40 percent of jobs in the region are exposed to the effects of AI, and that up to 17 million workers may not benefit from it due to a lack of digital infrastructure or adequate skills.
The paradox is that, in a continent that exports talent, the vast majority of its population continues to be passive users of technologies designed and governed by other regions. The dilemma then is not only economic or labor-related, but political. Who defines the rules of the game in this new industrial revolution? The technological corporations that develop the language models and control the data already set the pace of change, but States, universities and local communities seem to be running behind, trying to understand a phenomenon that advances faster than laws and budgets.
In the most advanced countries in the region, such as Chile, Brazil or Mexico, national artificial intelligence strategies have already been promoted. However, few include real governance or citizen participation mechanisms. Policies often focus on promoting technology adoption without asking who controls the algorithms, how personal data is protected, or what happens to displaced workers. In other words, we talk a lot about innovation, but little about technological evolution. The risk of digital dependence looms over Latin America: just as in the 20th century the region depended on the import of industrial machinery or military technology, today it could become trapped in the import of algorithms. We consume platforms that do not reflect our languages, our values, or our realities. If we do not develop our own models, artificial intelligence will end up being a new form of colonialism, one in which it is not territories that are dominated, but data.
More initiatives should be promoted, such as when, for example, in June 2025, several Latin American countries launched a joint initiative to develop a regional language model, Latam-GPT, aimed at incorporating expressions, accents and contexts typical of Latin American Spanish and Portuguese. This is not just a linguistic issue, but a cultural and political one, and an opportunity to encourage the possibility of our algorithms learning from us and not only from foreign databases.
If this experience prospers, it could mark the beginning of regional technological sovereignty, something that Latin America has not had for decades. But beyond innovation, the crucial debate remains social. In an economy where almost half of employment is informal, automation can be devastating if not accompanied by active policies. The most vulnerable sectors such as administrative services, retail, customer service, logistics, are among the most exposed.
In many cases, artificial intelligence replaces tasks, not entire people, but job fragmentation can erode the stability of millions of households. Without protection networks, professional reconversion or support for entrepreneurship, technologically displaced people will be condemned to precariousness. The other side of the problem is the invisible workers behind artificial intelligence, that is, those who label, correct or validate data from Latin America to feed global models.
Talents from Venezuela, Argentina, Ecuador, Brazil, among others, carry out this remote and precarious work for minimal salaries. It is the paradox of a region that, while discussing the future of work, already supports part of the global digital economy in 19th century conditions. If Latin America wants to transform this crisis into an opportunity, it must act now. The first step is to define national artificial intelligence policies with a social, not merely productive, focus. AI cannot be just a topic of business innovation, it must include mass training strategies, inclusive connectivity, technological education and ethical regulation. Without extensive digital literacy, the majority of the population will be trapped between fascination and fear.
Digital infrastructure continues to be the great bottleneck, since there is no inclusion without access, and without access there is no equity. Investments in fiber optics, 5G and satellite networks should be considered as strategic as road or energy works. Likewise, education needs a content revolution, aimed at training new generations in algorithmic thinking, data ethics and digital creativity, not only in the instrumental use of technology. Smaller countries, such as Ecuador or Uruguay, can find a comparative advantage in this transition if they opt for agile policies and alliances with universities, startups and local governments. Ecuador, for example, could promote an AI strategy applied to its structural challenges, such as water management, sustainable agriculture, disaster prevention or rural education. If the country manages to link technology with social purpose, it could position itself as a regional benchmark in inclusive innovation.
The ethical component is equally urgent. Citizens have the right to know when a public decision has been mediated by an algorithm, how their data is used, and what biases exist in the models that affect their lives. Algorithmic transparency and independent audits should be part of the democratic agenda. Artificial intelligence should not replace politics, but improve it. For these reasons, the great Latin American challenge is not in learning to use artificial intelligence, but in deciding what we want to do with it. If we allow innovation to advance without direction, we run the risk of deepening our inequalities, but if we transform it into a tool for human development, we could be facing a historic opportunity towards a productive and educational leap.
In this dilemma, the role of universities, scientists and public thinkers will be decisive. We need voices that connect technology with ethics, productivity with justice, innovation with empathy. The future of work cannot be designed from laboratories disconnected from social reality. Artificial intelligence confronts us with a moral and political question: do we want a technology that replaces us or one that complements us? The answer will depend on whether Latin America decides to be the author or simple spectator of its own future.