Hands-on-Data: Artificial Intelligence for the Design of Public Policy in Latin America


Contributed Panel in English. Although data produced by governments are abundant, in developing countries data-informed public policy design and implementation is often more an aspiration than a reality. The Hands-on-Data (HoD) initiative is a successful practical experience that gathers several stakeholders (e.g., public servants, academia- and industry-based data scientists, and a multilateral financing institution) to prototype data science products addressing specific public sector needs in Latin American countries. The first edition of HoD took place in Argentina during the last quarter of 2018, and resulted in five prototypes that used artificial intelligence (AI) and machine learning (ML) to address narrowly-defined public policy problems. Most importantly, HoD Argentina succeeded at organizing a seven-week taskforce consisting of public officials and data scientists, who extracted actionable knowledge from large datasets (e.g., administrative records) while ensuring anonymity and confidentiality of sensitive information. The policy domains analyzed in HoD Argentina were: work and labor, health, education, urban public transport, and accessibility in large cities. The short but intensive interaction between the public sector and data scientists sparked the interest of key stakeholders, and resulted in new co-creation initiatives that exceed the timeframe of HoD. In this way, HoD acted as a stepping stone for new data collaborations.

University College of London, London, United Kingdom