Measuring corruption

The use of big data analytics and artificial intelligence systems as tools for measuring corruption seems to confirm and strengthen two trends that had already emerged with the second-generation indicators. The first trend is to increasingly use administrative data as a source to learn about and try to better quantify the level of corruption in a state. The second trend is to move beyond the measurement of country-specific corruption and try to quantify it with an approach that pays more attention to specific sectors with a higher corruption risk, such as the public procurement sector, or to territorial differences in corruption risk within a state. These two trends pose a number of problems that need to be addressed in order to allow new technologies to express their full potential for the benefit of policy makers and society at large. Both issues need to be faced mainly within the responsibility of the administrative systems of European countries. The first regards the need to guarantee the availability, obtainability, usability and, in general, the quality of public data. The development of preventative indicators based on big data analytics and artificial intelligence systems, capable of exploiting all the reconstructive and predictive potential of these technologies, is only possible if public databases are open, well organized, complete and up to date. The openness of public data is then the first, basic condition to guarantee a positive cognitive return, useful to the administration system itself. The second issue concerns the capacity of European administrative systems to effectively employ and take advantage of these tools in control activities and in the planning of administrative corruption prevention policies. Considering that the objective of many European countries, such as Italy, is to move in this direction, the proper organization of public databases is a crucial issue. Equally, it is important to have an administrative organization that is functional to the use of new technologies, i.e., able to understand its own cognitive needs to fight corruption and, consequently, to build new collections of administrative data functional to these needs; able to equip itself with adequate technological tools and professional figures qualified to use them in support of policy-makers’ activities. Lastly, the regulatory issues that are still open at the European level in relation to the use of big data analytics and artificial intelligence systems during administrative activities are not marginal and need to be accounted for. The European General Data Protection Regulation, in particular, does not allow for a full exploitation of these tools, considering the limitation it imposes to massive data processing, especially when involving data of a personal nature, and decisions impacting identifiable subjects.