Leonardo Medrano, Secretary General of Academia y Desarrollo at Universidad Siglo 21, laid out a stark choice for educators this week: harness artificial intelligence to narrow educational inequality, or risk deepening it through unchecked deployment of powerful technologies that most students—and teachers—don't yet understand.
In a column published by LA NACION on October 5, 2026, Medrano didn't dismiss AI's potential. He named three ways it could serve learning, alongside three serious risks requiring an immediate response from institutions. The debate, he argued, can't stay stuck between techno-optimism and blanket restriction. Schools face a harder question: what should AI do, what should it never do, and what must students learn to live alongside it?
Where AI Could Close Learning Gaps
Today's education system relies on a troubling assumption: one teaching method, one pace, and one curriculum can work for students with wildly different backgrounds, struggles, and speeds. Medrano noted that modern education has dramatically expanded access to knowledge, but it has done so through "relatively standardized" models that ignore individual need.
AI could change that. A system can detect when a student needs more explanation, propose activities, adjust complexity, or suggest resources. For students moving faster, it can raise the challenge instead of making them wait. Medrano wrote, "A system can detect that a student needs more explanations about a concept, proposes new activities, adjusts the complexity or suggests other resources." Spotting struggles early matters, too. Student interactions with educational platforms produce rich data: participation patterns, completion rates, study times, and signs of procrastination. Educators can use this information to catch difficulties early and offer timely support.
Students could also gain genuine agency. Instead of passively receiving instruction, they could ask for another explanation, request examples, generate simulations, practice across scenarios, test consequences, get instant feedback, and choose to increase the difficulty. Learning becomes active exploration, not just consumption.
The Deception Problem
Medrano also identified three hazards that can't be wished away. First comes disinformation dressed in plausible language. AI systems can generate convincing responses that aren't true. On a screen, a false claim can look exactly like a true one. Medrano wrote, "It's crucial to develop the capacity to evaluate information, verify sources, contrast arguments and identify biases." Without direct instruction in these skills, students won't know how to challenge a machine's answer.
The second hazard is cognitive erosion. Less cognitive effort can weaken cognitive capacity. If AI answers every question instantly, creates unlimited practice problems, and explains concepts from endless angles, students may stop thinking for themselves. Medrano put the challenge this way: "The educational challenge consists in distinguishing between using artificial intelligence to enhance thinking and using it to replace it."
What Schools Must Actually Do
The third risk goes deeper: dehumanization. Technology can answer questions around the clock and provide hundreds of exercises. But education involves more than delivering information. Medrano was clear about what calls for human presence: "accompanying, challenging, inspiring, listening and building community." If AI takes on routine work, schools must protect time for these irreplaceable functions. They shouldn't use efficiency gains to cut staff or eliminate space for genuine human connection.
Medrano said students need to build competencies, not simply accumulate content. As information grows easier to access, "the value will be less in accumulating content and more in knowing how to use it." People will still need to explain ideas, build agreements, ask good questions, understand other perspectives, lead teams, and convey meaning. Machines won't do that well.
Creativity matters, too. When AI can produce a technically correct answer in seconds, people can contribute by asking unanswered questions, connecting separate ideas, imagining alternatives, and building new solutions. Schools need to cultivate that ability; they shouldn't assume it will develop by itself.
Medrano ended with a vision that moves beyond passive adaptation: "The future of education does not consist of accompanying a technological transformation, but of giving it direction." Using AI to expand learning capacity, creative power, understanding, and the ability to transform reality requires institutional choices. So does deliberately building safeguards against deception, dependence, and dehumanization.
Why This Matters:
Educational systems already face stark inequality. Students from wealthy backgrounds have access to tutoring, smaller classes, and individualized attention, while low-income families often contend with overcrowded classrooms and under-resourced schools. AI could personalize learning at scale and help identify struggling students early enough for educators to intervene. That outcome isn't automatic, though. Without strong public oversight, curriculum standards, and investment in teacher training, AI could deepen stratification: affluent families might get high-quality personalized learning while public systems use technology to cut costs and reduce human support.
Students also need explicit instruction in media literacy, critical evaluation, and creative thinking so they don't become passive consumers of AI-generated content. These skills call for careful teaching, not algorithmic substitution. Medrano's question—what to potentiate, avoid, and teach—is fundamentally about whether schools will use AI to deepen human learning and equity or outsource judgment to machines.