The Invisible Frontier
The country celebrates its digital progress while, with statistical discipline, it makes disappear those who never crossed the first frontier of written language. The next decade will confront us with a question the official narrative still does not dare to ask: what happens when a society builds its digital future on an alphabetic foundation that still does not…

The country celebrates its digital progress while, with statistical discipline, making vanish those who never crossed the first frontier of written language.
The next decade will confront us with a question the official narrative still does not dare to ask: what happens when a society builds its digital future on an alphabetic foundation that has yet to be consolidated?
The Dominican Republic is preparing for a technological transformation that its education indicators do not support. The government announced 6% illiteracy. UNESCO projects 4.5%. The 2022 Census hints at figures close to 11% when the right questions are cross-tabulated. Four official instruments from the same State, separated by up to six percentage points. The difference amounts to half a million people living in a statistical limbo: they exist for the purposes of the registry, but vanish when it is convenient to narrate progress. National strategies speak of sovereign artificial intelligence, semiconductor factories and global services as if that phantom half-generation were not there.
The problem goes beyond methodology. The breaks in the ENHOGAR series, the changes to the MICS model, population projections based on a census from fifteen years ago, the selective mortality of COVID-19 among older adults with the least schooling, the rural-to-urban migration that dissolves rural data into the urban average: all of these factors produce a cosmetic reduction of the indicator without any public policy having taught anyone to read. The statistic moves by demographic gravity, not by institutional will. Roughly RD$5,300 million spent between 2021 and 2024 on literacy programs certified, in their best year, 2,719 people. The unit cost —RD$1.48 million per adult who learned to read— exposes an apparatus that feeds a zero-sum mechanism.
This is where what I have been calling, for years, datacolonialism in its domestic version comes into play: a State that extracts statistics from its own citizens to produce narratives of progress, without giving citizens back the truth about themselves. Generative artificial intelligence works as a new layer mounted on top of pre-existing inequalities and amplifying them. Those who cannot read cannot ask. Those who cannot ask cannot iterate with a model. Those who cannot iterate are left outside the new productive regime. What is serious about the Dominican moment is that we are importing that AI layer before having built the alphabetic layer that gives it meaning.
We can also call this phenomenon the double-frontier trap. The first frontier is that of the alphabet: depending on which official source you choose, between 471,000 and 800,000 Dominican adults still cannot read this text. The second frontier —functional digital illiteracy— is forming right now, as I write. It belongs to those who can read but cannot tell synthetic content from authentic content. To those who can type but cannot formulate a prompt. To those who browse without understanding that their data is the currency they pay with. We have not yet closed the first frontier. The second has already run us over.
Symbolic industrialization reaches its most dangerous form here. The alphabetic foundation is built on our own ground, person by person, in schools that work, with teachers who know how to teach, with cohorts that finish what they started. MINERD's own APEA indicator says it without euphemisms: the Dominican average, adjusted for real learning, is equivalent to 5.48 years of schooling. That is the equivalent of having barely completed the first cycle of primary school. On that foundation we are planning to become a regional artificial intelligence hub.
The Inputs Trap —that we described in the Tabuga Think Tank— operates with the same logic in education as in technology. We import educational management models, digital platforms, internationally certified curricula, without building the endogenous capacity to evaluate whether they work. The consequence is predictable: we spend more, measure worse, communicate better. The cycle repeats itself.
There is a dynamic property of public data worth naming. Figures, thanks to digital diffusion, transform before they are destroyed. When a State omits or softens an indicator, that indicator reappears as a cost elsewhere in the system. Today's underreported illiteracy shows up tomorrow as digital exclusion, as disinformation fraud, as vulnerability to deepfakes, as superficial adoption of technologies that promise democratization and only deliver dependency. The statistical dynamic we boost in the short term we pay for, multiplied, in the medium term. Every saving has a deferred cost.
The next census will arrive in 2032. We have ten years of public policy ahead, built on a data point whose consolidation will take at least three more. Ten years in which every decision on educational investment, digital transformation, artificial intelligence regulation and cognitive sovereignty will be made blind. Ten years in which the frontier of the alphabet and the frontier of the pixel will overlap until they become indistinguishable for those on the wrong side of both.
The cognitive sovereignty I demand so often in these columns begins with a simple, almost premodern act: counting our own people properly. Knowing how many we are, how we read, what we understand, what we can decide for ourselves. Every digital strategy depends on that basic accounting. LatamGPT loses its point if we do not know how many Dominicans can put a question to it. The 4.0 industry collapses if half the workforce understands the world only through voice messages.
What is at stake goes beyond the exact percentage. The central point is recognizing that a State that does not dare to measure itself cannot manage its results either. A country that celebrates digital progress while hiding its alphabetic lag builds a two-story society with no elevator between the floors, and no fire escape.
The next decade demands that we document progress with precision. The knowledge base we leave the country will record the chips, the data centers and the laws we achieved. It will also record who we left behind while we advanced. That answer starts with a census, decisions, key actors and a State that prefers the discomfort of real data to the refuge of political calculation.
The invisible frontier is the most dangerous because you cannot see it until you are already on the other side.