Potere algoritmico, digital welfare state e garanzie per gli amministrati. I nodi ancora da sciogliere.

The public administration has long been invested by a process of digitisation, which has generated gains in efficiency within the offices and externally in favour of users. In recent years, however, the spread of the use of algorithms and, more generally, of new technologies that make use of big data has affected administrative action “from within”, changing the way decisions are made and public policies are developed, from transport to the health system, from public safety to welfare policies. Precisely in the field of social policy, the collection and processing of large amounts of data analysed and processed by algorithms is at the root of the rapid deployment of software and platforms used in predictive function for the release of benefits and performance, but also in function of control, verification and surveillance to prevent or sanction. The use of data, including personal data, by the administration in the adoption or denial of measures that grant benefits to the interested parties is not a recent phenomenon. The automation of this process, however, has changed its characteristics with undeniable benefits for citizens, for example, in terms of access to social benefits, uniformity of the relevant criteria and speeding up of practices, and for the administration in terms of the use of public resources and the prevention of fraud. The advent of digital welfare systems using algorithms and big data also increases the risks of discrimination and violation of fundamental rights, as recently noted in an UN report according to which in social protection services there are a significant lack of transparency in data processing systems through new technologies with consequences in terms of violation of the rights of interested parties (such as the right to privacy and the respect for private life), a lack of responsibility of the involved actors, public and private, and a high risk of errors and prejudices in the adoption of public policies and decisions. Starting from the analysis of a case study in the Netherlands, the work highlights the main problems posed by decisions through algorithms and analyses the remedies provided by the jurisprudence and the existing legislation, even at a supranational level. The paper concludes with some reflections on the critical issues still unresolved and on the effects that these may have on the administrative power and on the relations with the citizens.