El Salvador’s AI-Powered Education Pilot Delivers PISA Results Comparable to Germany and Sweden
A pilot program combining artificial intelligence, teacher training, connectivity and new learning methods has produced striking results in 171 public schools, offering a potential model for technology-driven education reform across Latin America.
SAN SALVADOR — El Salvador is emerging as an unexpected case study in the use of artificial intelligence to modernize public education after students participating in a government pilot program achieved results comparable with the average performance of students in Germany and Sweden in an international assessment.
The results, announced by the Salvadoran government in September, involved 1,198 students from 171 public schools participating in the country’s Education Modernization Project. The students were evaluated through PISA for Schools, an OECD-linked assessment designed to provide internationally comparable information about individual schools and their students.
According to the results presented by Salvadoran authorities, participating students performed above the country’s national average in reading, mathematics and science, with scores comparable to the averages recorded by Germany and Sweden in PISA 2022.
The results are particularly notable because the program incorporates artificial intelligence tutors, digital technology, connectivity, teacher and school-director training, new educational materials, learning monitoring and changes in classroom methodology.
AI as part of a broader education transformation
Salvadoran President Nayib Bukele has highlighted the use of AI tutors as one of the most innovative components of the program. The initiative has been operating in the 171 pilot schools for more than a year.
However, government officials have stressed that the results should not be attributed exclusively to artificial intelligence.
The pilot combines several interventions designed to change how students learn and how teachers monitor educational progress. These include technology infrastructure, pedagogical support, continuous assessment and professional development for teachers and school leaders.
This distinction is important. The PISA results demonstrate the performance of students participating in the program, but they do not constitute a controlled scientific experiment capable of isolating the individual effect of AI from the other components of the reform.
In other words, the Salvadoran experience is less about “AI replacing teachers” and more about using AI as one component of a broader technology-enabled education system.
From 171 schools to more than 1,000
The government says the initiative has already expanded beyond the original pilot and is now being implemented in more than 1,000 public schools, with the objective of extending the model to the entire public education network within approximately 18 months.
That ambition could make El Salvador one of Latin America’s most closely watched experiments in large-scale AI-supported public education.
The potential implications extend beyond the country itself. Many Latin American education systems face similar challenges: teacher shortages, uneven access to educational resources, differences in school infrastructure and connectivity, and difficulties providing individualized support to students.
AI-based tutoring systems could potentially help address some of these gaps by giving students access to personalized assistance while providing teachers with additional tools to monitor learning.
The numbers require an important qualification
Despite the striking comparison with Germany and Sweden, the results should be interpreted carefully.
The 171 schools were not selected as a statistically representative sample of El Salvador’s entire public education system. They were selected for conditions that made the initial implementation of the modernization program possible, including infrastructure, connectivity and the willingness of local school communities to participate.
The 1,198 students therefore represent the population evaluated within the pilot, rather than the Salvadoran education system as a whole.
This distinction has also been emphasized by Bukele himself. The president acknowledged that the results do not mean that every Salvadoran school has reached the educational level of Germany or Sweden.
Instead, the results demonstrate something potentially more important for policymakers: students in Salvadoran public schools can achieve internationally competitive results when provided with a different educational environment and the appropriate combination of technology, pedagogy and support.
An encouraging signal amid a global education challenge
The Salvadoran results arrive at a time when many education systems are struggling to reverse declines in student performance.
The latest international PISA results have shown significant deterioration in mathematics and reading performance across OECD countries over the past decade. Against that backdrop, the improvement observed within the Salvadoran pilot has attracted international attention.
CAF, the Development Bank of Latin America and the Caribbean, has also highlighted El Salvador’s modernization effort, noting that the PISA for Schools 2026 results from the 171 participating schools showed particularly strong outcomes in reading, mathematics and science compared with the country’s national results.
The next challenge will be determining whether these results can be reproduced when the model reaches schools with different levels of connectivity, infrastructure and socioeconomic conditions.
A potential model for Latin America — and beyond
For governments across Latin America, the Salvadoran experiment raises an important question: Can artificial intelligence help democratize access to high-quality education?
The answer will depend not only on the sophistication of the technology, but on how it is integrated into the education system.
Connectivity, teacher training, curriculum design, student assessment and institutional leadership remain fundamental. AI may provide a powerful additional layer, but technology alone is unlikely to transform educational outcomes.
El Salvador’s experience is therefore worth watching not simply as an AI story, but as a public-sector transformation experiment.
If the results observed in the initial 171 schools can be reproduced across thousands of schools, the country could provide Latin America with a valuable example of how digital infrastructure, artificial intelligence and human-centered education can work together at national scale.
For the Gulf states, which are investing heavily in artificial intelligence, digital government and future-ready human capital, the Salvadoran case also offers an interesting perspective: AI in education is moving from experimental classrooms toward large-scale public policy.
The coming years — and particularly future national assessments — will show whether the promising results of the pilot can become a sustainable transformation of El Salvador’s education system.
