Transparency requirements for algorithmic recommender systems, content selection and online advertising in the new European Digital Services Act

Abstract

This paper addresses the main issues raised by artificial intelligence systems, which are used by certain Internet service providers to personalise online advertising and to select and recommend information and content. It examines the obligations arising from the General Data Protection Regulation (GDPR), with particular attention to the principle of proactive liability and transparency, in relation to these systems. Finally, it deals with the new developments introduced by the new European Digital Services Act (DSA) in the field of transparency. In particular, the new algorithmic transparency and accountability obligations for very large online search engines and very large online platforms are studied, considering that the new Regulation offers useful tools and important guarantees to fight disinformation and protect the right to receive pluralistic information, freedom of expression and ideology, as well as to prevent the dissemination of illegal or harmful content.