On the path to the future: mapping the notion of transparency in the EU regulatory framework for AI

ABSTRACT Transparency is the currency of trust. It offers clarity and certainty. This is essential when dealing with intelligent systems which are increasingly making impactful decisions. Such decisions need to be sufficiently explained. With the goal of establishing ‘trustworthy AI’, the European Commission has recently published a legislative proposal for AI. However, there are important gaps in this framework which have not yet been addressed. This article identifies these gaps through a systematic overview of transparency considerations therein. Since transparency is an important means to improve procedural rights, this article argues that the AI Act should contain clear transparency obligations to avoid asymmetries and enable the explainability of automated decisions to those affected by them. The transparency framework in the proposed AI Act leaves open a risk of abuse by companies because their interests do not encompass considerations of AI systems’ ultimate impact on individuals. However, the dangers of keeping transparency as a value without a legal force justify further reflection when regulating AI systems in a way that aims to safeguard opposing interests. To this end, this article proposes inclusive co-regulation instead of self-regulation so that impacted individuals as well as innovators will be empowered to use and trust AI systems.