Artificial intelligence (AI) has positioned itself as a new hopeful tool that promises to help solve the problems of living in a disconnected, analog society. It pledges to transform society, the economy, and the system of governance for the benefit of citizens. How will it achieve this? That remains uncertain, but before worrying about those minor details, the focus seems to be on a near future where the digitization of information breaks centuries-old barriers and long-standing social disparities disappear under the power of a binary society.
However, on the flip side, led by technophobes and the occasional student of realpolitik, AI brings with it significant risks if not implemented with a perspective of digital equity. AI is just another tool accessed through the various telecommunications networks within a market, with the added complication that having the right device to interact with it will become increasingly important.
Given this reality, AI is in a growth phase in Mexico that could establish the country as a key player in Latin America. However, it must overcome structural obstacles threatening to widen the digital divide and reproduce existing inequalities to achieve this. If there is no telecommunications service, there is no AI; if service exists but is too expensive, there is still no AI; and even if both of these conditions are resolved, without the right device, it’s as if AI didn’t exist.
Nonetheless, compared to other Latin American countries, Mexico has clear advantages: its economic size, the breadth of its domestic market, and its commercial proximity to the United States have historically placed it in a privileged position. However, the development and adoption of AI may become one of the many collateral damages caused by the Trump administration’s failed tariff strategy toward Mexico.
Mexico has a growing tech ecosystem and prestigious universities such as UNAM and the Monterrey Institute of Technology (Tec de Monterrey), which, through initiatives like Microsoft’s LIAM and the Tec de Monterrey AI Hub, promote projects in natural language processing and other AI fields.
Still, a lack of robust infrastructure and a shortage of highly specialized talent are major challenges, as pointed out in reports like Oxford Insights’ Government AI Readiness Index, which places Mexico in the middle ranks of Latin America regarding AI preparedness in its 2022 edition. The lack of human capital and the brain drain of digitally skilled professionals is not just a Mexican problem—it’s one of the main hurdles across the region.
The immediate future of AI in Mexico is further complicated by a transitional regulatory framework that has only succeeded in uniting the market in opposition to the proposed telecommunications and broadcasting law. What is clear is that President Claudia Sheinbaum’s administration’s tendencies will likely lead to a legal framework for AI in which control and final decision-making in any crisis rest with the federal executive branch.
While the legal parameters for AI are still being debated, a large portion of Mexico’s population continues to experience high levels of poverty and inequality, according to data from the National Council for the Evaluation of Social Development Policy (CONEVAL). Figures from the National Institute of Statistics and Geography (INEGI) show that the areas with the lowest connectivity are rural and Indigenous regions. This limits the equitable adoption of emerging technologies like AI by the very people who could benefit from it as a tool to escape poverty.
One of the most relevant fields in AI today is the development of Large Language Models (LLMs). In Mexico, this area is still in its early stages. Most models come from major foreign tech companies like OpenAI, Google, or Meta. While there are local efforts by universities and research centers, these remain limited in scale.
Returning to the topic of the proposed legal framework, particularly Article 109, if all AI-related efforts are concentrated in foreign entities, how would the State react if, through one of the AI platforms operated by these companies, the responses to user queries were deemed by authorities to be offensive propaganda against the nation?
Regardless of how the law ultimately plays out, the biggest barrier to developing a Mexican and inclusive AI is restricted access to high-quality linguistic data in Mexican Spanish and Indigenous languages, along with a lack of the computational infrastructure needed to train homegrown models. Without a clear national strategy or meaningful incentives, the autonomous development of LLMs remains a distant goal.
Once that dilemma is resolved, the next step would be determining how to increase the generation of truthful data to feed AI. The effectiveness of AI models—and particularly LLMs—depends on the quality and representativeness of the data they are trained on. This gives a substantial advantage to regions with mature digital ecosystems and large volumes of data, like the United States or Europe, while putting countries like Mexico at a disadvantage unless public policies are implemented to promote the collection and sovereignty of local data. It’s also essential not to overlook the need to digitize key sources of information that provide historical context for present-day data.
Currently, the primary beneficiaries of AI are urban populations with higher socioeconomic status. Meanwhile, rural, Indigenous, or low-education communities risk being left behind. This exclusion is not just technical; it’s also cultural and economic, ranging from lacking multilingual interfaces to the concentration of benefits in large corporations. Without public intervention, there is a danger that AI will deepen existing inequalities rather than remedy them.
Part of the solution has been proposed for decades: provide broad telecommunications coverage, offer service plans and devices at affordable prices, and educate people to recognize the value of telecommunications. Hence, they see it as a worthwhile investment. Without addressing these three areas, AI will continue to appear on magazine covers, in political speeches, and in movie plots based on Isaac Asimov’s stories.
But let’s remain optimistic: Mexico has clear opportunities despite these challenges. The country boasts a rich linguistic base, an active scientific community, and economic sectors with high potential to benefit from AI. If President Sheinbaum’s administration strengthens digital infrastructure, establishes a clear regulatory framework, and promotes public-private partnerships aimed at responsible development, Mexico could become a regional leader in AI and close, rather than widen, its digital divide.
Let’s hope it succeeds, because if it does, Mexico stands to benefit the most.
