evidoria

← Back to browse

Good practice Imported

Plan Ceibal PAM — Uruguay’s AI adaptive mathematics platform for public schools

Uruguay · Montevideo · See the Uruguay profile · See the Montevideo profile

Evidence: Quasi-experimental Top 35% 67/100 · Ask Evidence Copilot about this practice

Uruguay’s Plan Ceibal deploys PAM, an AI-driven adaptive maths platform, to 100% of public schools (3,038 institutions, 733,000 students). An independent evaluation found intensive PAM use raised mathematics attainment by 0.2 standard deviations; the IDB has since financed expans

0.2 standard deviations
Increase in primary-school mathematics attainment for intensive PAM users vs. matched control groups (2017-2019 evaluation)
3,038 schools (100% of public institutions)
Public schools with PAM access (December 2021)
733,000
Enrolled students with PAM access (December 2021)
100,000+
Exercises in PAM's adaptive learning library

Details

Maturity
Established
Promoter
Plan Ceibal
Period
2013–present
Keywords
adaptive learning, mathematics education, machine learning, public education technology

Context

Plan Ceibal is Uruguay's state programme for educational technology, operating under the national education authority (ANEP). Since 2013 it has offered licences to PAM (Plataforma Adaptativa de Matemáticas), an adaptive learning platform that uses machine-learning algorithms to personalise mathematics instruction across the full primary and secondary curriculum.

Objectives

Personalise mathematics instruction for each student based on their response patterns, including the type and frequency of errors, and give teachers real-time dashboards for targeted intervention rather than one-size-fits-all instruction.

Activities

PAM recommends a dynamic learning path from a library of more than 100,000 exercises. By December 2021, PAM was available to 100% of Uruguay's public educational institutions (3,038 schools) and their 733,000 enrolled students, alongside universal broadband connectivity in schools. The Inter-American Development Bank subsequently financed Plan Ceibal II to expand PAM coverage further and assess long-term impacts.

Results

An independent economic evaluation by Perera and Aboal (CINVE / IDB, 2017-2019) found that intensive use of PAM increased primary-school mathematics attainment by 0.2 standard deviations compared with matched control groups.

Conclusions

The model has been studied internationally by the World Bank and the Global Partnership for Education as a replicable approach to technology-enabled personalised learning in public education systems. A cautionary note: PAM relies on Bettermarks, a commercial adaptive platform, raising questions about long-term data sovereignty and algorithmic transparency.

Implementation
Indicative cost
High (€500k–€5M) — Not itemised in the source; involves nationwide licensing of a commercial adaptive-learning platform (Bettermarks) plus IDB-financed expansion (Plan Ceibal II) across 3,038 schools.
Time to results
Long (> 3 years) — Licences offered since 2013; achieved 100% school coverage by December 2021; IDB financing continues expansion and long-term impact assessment under Plan Ceibal II.
Staffing & skills
Plan Ceibal / ANEP programme staff, Teachers using real-time performance dashboards for targeted intervention

Conditions for success

  • Universal school broadband connectivity (achieved by December 2021)
  • Teacher-facing dashboards enabling targeted intervention
  • Independent impact-evaluation partnership with CINVE/IDB

Common failure modes

  • Reliance on Bettermarks, a commercial adaptive-learning engine, limits algorithmic transparency and raises data-sovereignty questions
  • No documented algorithmic redress mechanism for students

Where it fits

Governance type
national state education programme under the Ministry of Education (ANEP)
Scale
national — 100% of public schools
Income level
upper-middle-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful