The democratic credentials of non-state actors in the global governance of artificial intelligence

In the last decades, non-state actors have been ascribed a central the role in the democratization of global governance. The hope is that non-state actors representing civil society, such as NGOs, advocacy groups and social movements, can help address a democratic deficit in global governance, by promoting increased inclusion, representation, transparency and openness towards citizens around the world. International organizations in several policy areas, such as global environmental governance and global health governance have expanded their existing interaction with civil society, and organizations that historically have been rather closed have moved towards greater openness. A similar tendency can be observed in a relatively new policy area which has been growing exponentially in the last couple of years, namely, the global governance of artificial intelligence (AI). What is particularly interesting in the AI domain is that the non-state actors engaged include not only civil society actors but also, and primarily, the very same multinational tech companies that develop the AI technology that is to be regulated, and which possess large economic resources, power, and expertise concerning AI technology. Despite the fact that almost 600 governance initiatives have been launched globally since 2017, many of which authored by non-state actors, largely no attempt has been made to theorize the normative requirements of these actors from a democratic point of view. To fill this void, this paper analyzes the democratic credentials of non-state actors in global AI governance. Adopting an ecumenical understanding of democracy – which assumes that governance arrangements are democratic to the extent that those affected by the decisions have a direct or indirect say in the decision-making on equal terms – we argue that non-state actors can play an important role in global AI governance from a democratic standpoint, exercising what we describe as moral, epistemic, or market authority. More specifically, by introducing a distinction between the notions of ‘democratic agent’ and ‘agent of democracy’, our twofold aim is to (a) specify under what conditions non-state actors contribute to the democratization of the global governance of AI (as democratic agents) and (b) under what conditions they rather contribute to the strengthening of the prerequisites for democratization (as agents of democracy).