The unprecedented pace of innovations and the widespread adoption of information and communication technologies (ICTs) worldwide have given rise to a unique set of opportunities for governance. Governments have harnessed these technologies to streamline processes and enhance online access to public goods and services. Concurrently, citizens have benefited from increased internet accessibility, allowing them to access these services and stay informed about their leaders' activities and performance. However, alongside these promises, ICTs also present significant societal challenges for governance. The extensive sharing of information online, combined with inadequate quality checks, has given rise to a proliferation of misinformation. This misinformation has been linked to a decline in trust in institutions and science, as well as an increase in polarization, xenophobia, and foreign interference in the political sphere. Similarly, the lack of transparency and oversight in technologies like artificial intelligence and machine learning has raised concerns among experts and policymakers regarding their potential misuse. Confronted with these challenges, the international community has called for widespread digital literacy training to maximize the benefits of ICTs while mitigating the risks mentioned above. The "DIGILITERACY" initiative aims to explore the potential of digital literacy to enhance the economic and social well-being of both advanced and developing countries that have implemented this policy. In particular, it seeks to understand the distinct effects of this policy on two key stakeholders in the governance sphere that can disproportionately benefit from it: civil society and bureaucrats. The project will employ state-of-the-art quasi-experimental and experimental methods to assess its causal effects and evaluate its impact on digital media skills, misinformation consumption, political polarization, and the provision of public goods, all of which will be measured using a combination of originally collected and administrative data.

The objective of this project is to audit machine learning algorithms and make them compliant. The new legislation (RGPD and European Act) provides a legal framework that will frame the practical implementation of algorithms. They provide a number of recommendations that algorithms must follow. In particular, these algorithms must not behave differently for sub-groups of users unless these sub-groups are identified in advance and the differences are justified. They should also clearly display what they are designed to do and not mislead users. A large body of research exists to assess bias in machine learning as well as to study the explainability of algorithmic decisions. In the first place, this work should be pursued in order to better understand these problems and to develop procedures to certify the presence or absence of bias. In addition, the difficulty of auditing algorithms comes essentially from the fact that the measurements depend on the distribution of the sample. But in a "black box" auditing framework, i.e. knowing only the outputs of the algorithm on a data set previously selected or chosen by the auditor, it is necessary to take into account the variability of the algorithm with respect to the distributions themselves. Our objective in this project is therefore to develop new ways of defining, detecting and controlling the effects of biases, in a uniform and robust way when the distribution of the observations is partially known. Our approach is multi-disciplinary, relying on robust statistics and machine learning (maths and computer science) to define valid properties for distributional neighbourhoods, Gaussian processes for the construction of optimal experimental designs for the discovery of observations, and optimisation to be able to build algorithms practically.

School choice programs allow students (and their parents) to select a school of their preference in an attempt to improve access to high-quality education. As school capacities are limited, it is not possible to assign every student to their most preferred choice. Therefore, it is important to define a mechanism for student assignments. There is little evidence on the impact that these mechanisms have on equal access to quality education. The way in which they are implemented could hurt socially disadvantaged groups for several reasons. For example, it could reduce the options available for disadvantaged students if schools are able to select students. Moreover, misinformation about the rules, or limited resources to apply and hold out for the best options, could disproportionally harm students from low socioeconomic backgrounds. These are featured frequently across many different school choice settings without a clear understanding of the consequences in terms of educational outcomes and inequality. Our team proposes a project in which we will analyze education programs that aim to improve outcomes among the socially disadvantaged. We will study how the characteristics of the allocation system impact educational outcomes for different socioeconomic groups both in compulsory and higher education. We will use and extend structural econometric models to study student behavior, which will allow us to study some of the determinants of social inequality and to simulate outcomes under alternative matching policies.