Artificial Intelligence in combating school dropout: The student retention diagnostic system (SDPA) in Brazilian basic education

• In Event: Highlighted Session: Technology, AI, and Policy in Education

Proposal

School dropout is a complex, multidimensional phenomenon with serious academic, social, and economic ramifications. In the Brazilian context, this issue extends beyond classroom boundaries, driven by a variety of interrelated factors. Among these, notable elements include violence, family environment instability, the need for early entry into the labor market, and deficiencies in the quality of education (Borja & Martins, 2014; Dore, Araújo & Mendes, 2014; Silva Filho & Araújo, 2017). It is therefore crucial to distinguish between temporary withdrawal—referring to students who interrupt their studies but return in the subsequent school year—and dropout, which denotes the definitive disengagement from school (INEP, 2013).

To address this challenge, the Brazilian State holds the constitutional responsibility of ensuring equitable conditions for access to and retention in compulsory, tuition-free basic education for children and adolescents aged 4 to 17 (Brazil, 1988).

Within this framework, the present text outlines an innovative research project currently under development at the Education Department of the Catholic University of Brasília (UCB). Its goal is the design and implementation of the Student Retention Diagnostic System (SDPA), a predictive diagnostic tool aimed at mitigating school dropout. Supported by funding from the Federal District Research Support Foundation (FAPDF), this initiative seeks to create an Artificial Intelligence–based prototype capable of proactively monitoring and intervening in students’ educational trajectories, thus fostering continuity in their academic paths.

In 2023, the Brazilian educational landscape recorded a total of 47.3 million enrollments across 178,500 basic education schools. Of this total, nearly half of students (49.3%) were enrolled in municipal schools, while the private sector accounted for 19.9%. The federal government’s participation in basic education remains minimal, representing less than 1% of total enrollments (INEP, 2024).

Data on educational setbacks—which include grade repetition and student withdrawal—between 2019 and 2022 underscore the urgency of the situation and the necessity of tools such as SDPA, which enable early identification and preventive intervention.

The SDPA has been designed as a tool for principals, teachers, and school supervisors. Its core functionality resides in the ability to compile, organize, and analyze data at the levels of schools, classes, and individual students. It encompasses metrics such as national standardized test scores (SAEB), age–grade distortion, school attendance, and dropout risk. Beyond facilitating student data management, the platform enables communication with parents and guardians and the generation of detailed reports.
Artificial Intelligence is applied in the SDPA to assess dropout risk (low, medium, or high) at the school, class, and student levels. This assessment is based on variables including attendance frequency, disciplinary records, and academic performance. Additionally, the system integrates socioeconomic indicators, such as social vulnerability and access to basic services. Risk levels are classified as follows: low risk, when dropout is projected to be under 20%; medium risk, with rates between 20% and 49%; and high risk, with rates equal to or exceeding 50%. Furthermore, the system incorporates an alert mechanism that notifies users regarding relevant updates.

The UCB Predictive Diagnostic Evaluation research project represents an effort to confront the problem of school dropout in Brazil. By leveraging the potential of Artificial Intelligence for monitoring and predicting the risk of student withdrawal, the SDPA emerges as a valuable resource for educational management, enabling timely and effective interventions.

It is understood that the successful implementation of systems such as SDPA carries the potential to positively impact student retention and enhance the quality of basic education, thus fulfilling the State’s responsibility to guarantee educational permanence. In this perspective, the project exemplifies how Artificial Intelligence can serve as a key ally in addressing social and educational challenges, helping to combat a phenomenon with long-term negative consequences for society as a whole.

Authors

• Renato Brito, Catholic University of Brasilia

• Maria Cristina Mesquita da Silva, Universidade Católica de Brasília