Artificial Intelligence in Education: Learning by Doing and Evaluating to Scale

For decades, Latin America has invested expectations — and resources — in educational technology. Results have been uneven, as the real benefits of technological innovation depend on several concurrent factors, including strategic leadership, integration into the curriculum and school practice, the level of digital training and literacy among teachers and students, and the assurance of adequate data and technology infrastructure.

The recent evolution of artificial intelligence (AI), especially generative AI, presents a real opportunity to transform practices and improve educational outcomes — but also carries the risk of repeating old mistakes if adopted uncritically. From the perspective of schools, what can make a difference is having an approach grounded in explicit pedagogical objectives and processes; impact evaluation of innovation experiences from day one; prior preparation of teaching staff; and careful implementation that guarantees equity. This was made evident at the recent event "Effective Learning with AI in Literacy and Mathematics," held in Buenos Aires, Argentina, on October 21, 2025, jointly organized by the IDB and Universidad Austral, with the support of the National Secretariat of Education.

Objectives and Enabling Conditions

Latin America and the Caribbean faces four challenges that AI can help address if integrated with pedagogical purpose and clear rules.

1. Low literacy levels and learning gaps that demand early recovery and acceleration.
2. Interrupted educational trajectories and dropout risk, whose timely detection and response require systems based on individual-level data.
3. Teaching competencies in general and digital literacy in particular, which can be strengthened through AI adoption.
4. Management efficiency and resource allocation, so that support reaches those who need it most.

This roadmap does not emerge from technology itself, but from leaders' capacity to identify and characterize problems. This was one of the central themes of the event, which repeatedly emphasized the need for a careful, evidence-based implementation of AI in education — with clear objectives and a framework that prioritizes freedom, ethics, and the development of local capacities. It also underscored the opportunity to anchor these initiatives in national agendas such as the National Literacy Plan, in order to align efforts and resources within agreed goals and coordinated policies.

The event provided a space to reflect on key findings from the IDB publication "Artificial Intelligence and Education: Building the Future through Digital Transformation,"whose authors warned against the risk of overestimating the impact of "the digital" without sufficiently securing its pedagogical integration and infrastructure conditions. Beyond acknowledging AI's potential contribution, participants cautioned against repeating that pattern if governance, evaluation, and monitoring aspects are not addressed from the outset.

Lessons from On-the-Ground Implementation

Presentations of regional and provincial experiences highlighted the efforts underway across different parts of the country and the region to introduce "purposeful AI" into day-to-day management.

Mendoza, for example, integrated an Early Warning System (EWS) into GEM — its provincial management platform — which uses AI models to anticipate risk levels and categorize students in real time based on attendance, performance, participation, and context. That categorization is not an end in itself, but rather a trigger for the use of dashboards that enable timely interventions in schools. The province also accompanied the rollout with teacher professionalization to sustain pedagogical use; the focus, again, was on the educational response — not on technological novelty.

Santa Fe advanced in the same direction with a predictive model drawing on fifteen years of individual-level data to strengthen its EWS, identify at-risk students early, and design targeted responses. The implementation lesson is twofold: without a data culture sustained over time — technical teams, procedures, interoperability — AI delivers no value; with those capacities, however, it enables more precise prioritization and resource allocation.

Both cases confirm a guiding principle that the National Secretariat of Education team framed through a humanist lens: educational innovation first, then "with AI." Technology is a means, not an end; the teacher remains the agent of change, and the school is the space where an alert becomes action. A human-centered narrative was recognized as a key anchor for aligning stakeholders and giving coherence to initiatives — especially when technological uncertainty fuels outsized expectations. In this context, the Argentine Program for Educational Innovation with Artificial Intelligence (PAIDEIA) was presented as an initiative aimed at bringing coherence to the incorporation of AI in the education system, through a comprehensive lens that cuts across levels and dimensions. It was emphasized that AI is conceived as part of a broader educational transformation.

Governance and Scale: The Path toward Sustainable AI and Education Policies

The presentation highlighted four cross-cutting pillars — student support, teacher training, content generation, and integration into management systems — alongside an interdisciplinary and cross-sectoral approach that combines technical, pedagogical, ethical, and social knowledge. From this perspective, it was underscored that education remains a profoundly human experience and that AI must be placed at the service of relationships, dialogue, and situated pedagogical decisions.

Against this backdrop, it became clear that Argentina offers fertile ground for learning by doing with AI and measuring in order to scale. Timely in this regard was the publication of the report "Literacy + AI: Challenges and Opportunities in Argentina," which systematizes the exchanges from the previous edition of the AI and Literacy gathering held in 2024. The report highlights AI's potential as a tool for personalizing learning, automating processes, and supporting pedagogical decisions — but with equal force warns of the risk of widening the digital divide, and stresses the need to prioritize teacher training and ensure data governance.

The conclusion of this new event was clear: generating local evidence before scaling is essential. This is particularly important in the context of implementing the National Literacy Plan — agreed upon by all jurisdictions — which provides an important framework for guiding efforts to test AI resources that strengthen reading and writing and the monitoring of student trajectories, using common criteria and comparable metrics. In this regard, PAIDEIA is advancing the integration of a reading fluency measurement tool in coordination with jurisdictions, aimed at producing student performance indicators that can inform potential pedagogical interventions and contribute to the development of evidence-based education policies.

The 2025 Buenos Aires event continued a dialogue initiated weeks earlier in a neighboring country. The Regional Congress on Artificial Intelligence in Education in Montevideo brought together ministries and representatives from Ceibal, UNESCO, the IDB, the World Bank, and research teams, with an agenda that promoted reflection on "human by design." The meeting introduced key research questions related, among other things, to AI-supported teacher accompaniment and the challenges of regulation and financing. This conversation suggests that the region is already operating with explicit success criteria: effect and cost-per-student thresholds, implementation fidelity, data protection, and operational sustainability. To the extent that governments and schools orient themselves by these criteria, AI will cease to be a promising technology and become a useful instrument for implementing targeted education policy that drives measurable improvements in learning, trajectories, and efficiency.

In sum, the challenge is not to have more AI, but to demonstrate that — given the right enabling conditions and already-validated strategies — AI contributes to improving learning outcomes, making processes more efficient, and advancing toward complete and more equitable educational trajectories.