Antivirus para la Deserción — AI-Assisted Early-Warning System for University Dropout Risk in Medellín, Colombia

Fundación Antivirus para la Deserción began in 2019 within the Systems Engineering programme at the Universidad de Antioquia in Medellín, later extending to the National University of Colombia. It combines anomaly-detection and forecasting models with sociovocational, socioemotional, socioeconomic and academic accompaniment to identify students at elevated risk of leaving higher education, a problem affecting roughly half of Colombian university entrants.

The development team, 40% women, works with partnerships including the Colombian Ministry of ICT and the RUAV network. The Inter-American Development Bank's fAIrLAC programme documented the initiative as a case study in responsible learning analytics.

Both the IADB fAIrLAC case study and a Federación Antioqueña de ONG profile note real operational deployment across two university programmes, but neither publishes a rigorous, peer-reviewed effect size for dropout reduction — the evidence is organisational and testimonial rather than experimental.

The foundation itself names data access, data quality and cultural resistance from university stakeholders as its three biggest implementation bottlenecks, and stresses that any learning-analytics system needs a human-centred approach and formal data governance to be trustworthy at scale.