Cecilia María Vélez White
Exministra de Educación de Colombia
In Wisconsin (United States), Victoria (Australia) and the Buenos Aires Province (Argentina), artificial intelligence (AI) has been used to assist in the development of early detection models for school drop-out risks. The analysis of these three cases illustrates how the use of this emerging technology can help improve student attendance during the course of their schooling.
The Wisconsin project is the most comprehensive, since it has been operating for the longest period of time (since 2012). It produces impact results on the variables that it intended to intervene in and overcomes some of the difficulties that had arisen in the use of early drop-out detection models. It also allows subgroup analysis, and in the case of Milwaukee, student tracking in higher education. It has also assisted institutions in designing policies to address problems detected from the information gathered.
The Drop Out Early Warning System (DEWS) model is part of a broader project on the use of data in the education system (WISE). This complete information system has been made available to all members of the education system, through a portal where multiple variables are presented, along with various consultation levels which are applicable to different demographics. This facilitates wide dissemination within a data protection framework.
The way in which the WISE project supports educational administrators and institutions in the use of the information is especially important. Educational planning is thereby able to take into account the information provided by the system in determining actions to be taken with regard to educational problems. In much the same way, this use of information allows institutions to establish more personalized processes based on student characteristics, and even to design more personalized responses.
The cases of Victoria and the Buenos Aires Province
Analysis of the Victoria case in Australia called Student Mapping Tool (SMT) illustrates how a very simple tool provides educational institutions with better student information to prevent dropouts. This tool reinforces the design of actions taken to prevent drop outs, since these are included in the same system, thereby permitting tracking and evaluation. However, this interesting experience was only of short duration which highlights the conditions required for continuity of similar projects, such as the need for administrator involvement in the project and to establish incentives for use by educational institutions.
The case of Argentina, which was designed and implemented in the Buenos Aires Province, illustrates how establishing models based on a short series may be possible. As in the Australian case, it indicates the importance of ensuring a sufficient period of time for full implementation of the project and the need to provide incentives for its use by schools.
About the study
These three cases were analyzed in a study to be published as part of a CAF report on the strategic use of data and artificial intelligence (AI) in the public sector: ExperiencIA; Data and Artificial Intelligence in the Public Sector in Latin America. This study aimed to document project implementation cases that used data and Artificial Intelligence (AI) in the education sector at an international level for the design, implementation and monitoring of public policies, in the proposal of recommendations on the possible use of this technology by education sector administrators in Latin America and that may help them to improve their efficiency and effectiveness.
The work began from a review of the state of the art in terms of data and AI usage in the education sector. From there, it was concluded that data analytics for learning developed in recent decades is applicable in the search tools that allow personalization of teaching. This is because the generation of greater knowledge and monitoring of individual students enables the design of customized learning solutions. The tools used have focused on early warnings and the identification of interventions, but they do not yet propose actions to be undertaken for this purpose, so it has been very difficult to measure their impact.
Likewise, it describes limitations in these drop-out early warning systems related to functionality, student tracking outside of school, as well as the possibility of using more sophisticated models that allow the identification of specific subgroups.
Experiences of overcoming these limitations were investigated. They look to leverage the use of AI in handling school dropouts, identifying students at risk, taking measures to keep them in school and supporting the formulation of strategies for their long-term attendance in the educational system.
This study was carried out within a regional initiative framework of the State Digital Innovation Directorate of the Vice Presidency of Knowledge, CAF – development bank of Latin America –, which promotes the strategic use of data and artificial intelligence (AI) by the public sector in Latin American countries, in order to create social and economic value.
On September 15, the ExperiencIA, data and artificial intelligence in the public sector conference will be held , where this and other examples will be discussed in depth, alongside executives responsible for related areas.
